Peter Fleischer: Privacy...?

The polar bears committing to protect baby seals

2026-02-02T22:38:00.000+01:00

Everyone wants to protect kids online. Recently, Australia passed a ban on social media use by under-16’s. France is in the process of doing the same for under-15’s. And the UK is consulting on a similar law. We all know that online platforms have become (often knowingly, intentionally) a snakepit of dangers for kids, a ‘marketplace for predators”, and I fully support laws trying to protect children. But, unfortunately, the road to hell is often paved with good intentions, and these laws are a catastrophe for privacy…why?

The problem for privacy is simple: how does a platform identify who is a kid and who is an adult? There are two potential methods for such age verification. One, is the real-world method, such as demanding that users provide documentation to prove they’re adults, like credit card or driver license numbers. Platforms don’t want to use the real-world method, because they know that many of their users won’t provide them. These users simply don’t trust the platforms with that sensitive information. The platforms know that requiring real-world age verification will result in losing users, and these platforms really, really do not want to see their user numbers go down. Their stock prices are all premised on rising user numbers.

So, these platforms are all turning to the second method of age verification: the inferential method, using tracking and surveillance to “guess” which users are kids. These platforms track all their users, monitoring everything they do, read, post, click on, share, with the unblinking non-stop surveillance of machines that are then using AI analysis to infer who’s a kid, based on patterns of usage that the AI algorithms decide are typical of kids. Ah, shucks, the platforms now have to monitor everyone, all the time, including all adults, in order to guess which users are kids. For platforms that have business models based on monitoring, profiling and targeting users, these new legal obligations to “protect kids” are a blessing in disguise.

TikTok explained this privacy-invasive approach. The New York Post reported: “Amid government-sponsored social media clampdowns, TikTok is rolling out a new AI-powered age-detection tech across Europe that allows it to identify and purge accounts belonging to kids under the age of 13.

The new AI age control system analyzes profile info, posted clips, and posting patterns to determine if an account belongs to someone underage, the tech firm declared in a press blog post.” I would wager that TikTok is collecting and monitoring much more data than what they disclosed.

To understand all this, I traveled to Greenland, to meet a group of polar bears who formed an alliance to protect baby seals. The bears all had Greenlandic names, but showing their great sense of Arctic humor, they nicknamed themselves TikTok, Instagram and YouTube. When I first met the bears, I was terrified by their size. After all, they are the apex predators of their ecosystem. But they explained that they were committed to protecting baby seals. I asked them how they planned to do that. They explained that they would create a system to track, geolocate, monitor and profile every baby seal in the Arctic. All this data would be shared amongst all polar bears, who committed to use it only for the seals’ protection. But I asked: don’t you eat the baby seals, isn’t that your business model, and wouldn’t a predator love to track and monitor its prey? The bears glared at me, and for a moment, I thought they might crush my skull in their jaws. But then they relaxed, and said that, well, yes, some baby seals might not get protected, but some would, and I should just trust that they were responsible bears. After all, they guffawed, we’re tracking you too.

Santa’s Naughty and Nice list for privacy 2025

2026-01-05T11:54:00.000+01:00

Santa has been an early adopter of large-scale data mining and AI predictive algorithms to compile this year’s list of Naughty and Nice. A (naughty) elf secretly shared a copy with me. Santa is an older North European guy, who’s not afraid to call out the naughty ones, and he knows he’s not particularly popular in the world of Big Tech, especially since Santa doesn’t deliver 100$ million support yachts on the Silicon Valley wish list. Santa knows 2025 has been a terrible, terrible year for online privacy.

Naughty

The AI industry: engineered the biggest data theft in human history to train its AI models, running roughshod over intellectual property, copyright and personal data rights in the content it just took. It just took all that data, without anyone’s consent, without paying for it, and without any transparency about what it was taking. And despite a lawsuit here or there, the industry seems to be getting away with it. But Santa saw it. AI-generated slop and AI scams started infiltrating every aspect of our lives, with things getting worse, and the AI industry that made the tools to create this slop and scams washed its hands of the damage it was creating. Even AI companies have trouble distinguishing real from fake, good luck to the rest of us. But Santa isn’t resentful, and he gave the AI tech gurus a free one week vacation to learn AI scam techniques in Nigeria, and their teenage daughters are particularly welcome too.

Santa knows that AI models need to be trained on data, like human brains, learning, e.g., to associate the word “zebra” with a picture of a zebra. Neither the machines nor the humans need to have actually seen a living zebra to learn the association. But to enable the current state of generative AI, the AI models must be trained on vast amounts of data. Santa knows there are legal ways for AI companies to get this data to train their models.

Santa knows that nice AI players could license curated data sets: some companies have built businesses that build data sets specifically to sell to AI companies to train their models. Imagine a company building a vast library of cards, combining the picture of a zebra with the word zebra. Scale AI is an example of such data annotation companies. They build their databases by employing lots of very low-paid humans, and then license their database to the AI industry. Privacy laws around the world would permit the database annotations to label public figures, like Bill Clinton, in their database, but not private citizens, like most of us.

Santa also knows that AI players could license other people’s copyright data: companies that own datasets, in other words, own the copyright on their datasets, can license (sell) them. For example, The New York Times could license its archive of articles. But AI companies have in the past just taken the data, and sometimes they’ve been sued for it. One of them, Anthropic, settled with copyright owners for a very large amount of money.

https://www.reuters.com/legal/government/new-york-times-reporter-sues-google-xai-openai-over-chatbot-training-2025-12-22/

This lawsuit, and settlement, was based on copyright. But what about privacy? Santa doesn’t like when naughty people steal someone’s property. Santa has never seen an AI company pay individual humans for taking their personal data to train their models.

The online ads industry: a giant ecosystem of publishers, advertisers and online ad exchanges continue to develop ever more intrusive and secretive data collection, monitoring, profiling and targeting methods. The industry increasingly uses “fingerprinting” to uniquely identify you and me and everyone online. For example, I clicked on some random news website called Daily Mail: it informed me that I had the choice of “subscribing, paying money to Daily Mail” or alternatively to “consent” to sharing my data (basically everything it was technically able to collect) with its 1436 partners for profiling and targeted ads…and that I was invited to visit the websites of each of those 1436 partners to understand what they would do with my data. (This absurd farce passes as “consent” to the processing of my personal data, in the grotesque world of online advertising.) The online ads industry has become a giant data orgy, and guess the role that you and I play in that orgy. Santa offered the online industry honchos a free one week vacation to Jeffrey Epstein’s Caribbean island, to join a seminar on the legal concept of “consent” in the context of a “deal” between rich and powerful older men claiming to obtain the consent of much younger women and girls for intimacy. The online ads world repeats that pattern, where rich and powerful players claim to have obtained our “consent” to share and abuse our personal data with all their thousands of partners.

Third-party cookies: everyone’s favorite online tracking tech was on death’s watch as a privacy-invasive untenable technology, until Google decided to cease its efforts to phase them out. Google abandoned its project to improve privacy online. https://www.theverge.com/news/653964/google-privacy-sandbox-plans-scrapped-third-party-cookies Why? I’m guessing, because the ecosystem of publishers, advertisers and ad exchanges just enjoyed the current data orgy too much. Too bad for individual humans whose data will continue to be shared (or abused) amongst thousands of “partners”.

E.xtremely L.oud O.bnoxious N.arcissist tech billionaires: Santa just wants them to STFU. Santa offered E.L.O.N.s a free one week vacation to Nelson Mandela’s former cell on Robben Island, in solitary confinement. Santa also invited them to consider moving, or returning to, South Africa, permanently, if they want to escape California’s proposed “billionaire tax”.

Nice

The Regulators, in particular, the Irish Data Protection Commission continued its pretense to regulate privacy on behalf of all of Europe in its role as the “one stop shop”...flop…and successfully finished another year doing nothing to upset the big tech geese laying the golden eggs in Dublin. You can’t be nicer than that to big tech. Its sister regulator in the US, the Federal Trade Commission, spent the year doing even less, in a Trump/Musk-induced organizational coma, after Trump fired their Democratic commissioners. The competition authorities had even less impact: the most important antitrust case of the last quarter century (Google) ended with judge-imposed remedies that were so gentle that Google’s stock rocketed 9% on the news. Santa offered the regulators a one week vacation to a Buddhist retreat: the world will end, you can't do anything about it, just accept it.

Individual humans are starting to wake up and fight back to preserve their world: real people are waking up to real threats to their freedoms and privacy. Real people are starting to object to building large industrial data centers in their backyards. Real people are starting to object to AI models taking their data to train their models, or deepfake their voices or faces, or to make basic human decisions about their lives, like getting hired or fired, or scamming them. Real people are starting to object to the mass surveillance system that the open web has become. Santa offered them heart, brain, and courage…just like the Wizard of Oz.

How about you?: have you been naughty or nice this year? Santa knows the naughty ones have gotten much, much richer this year, because the other 8 billion people got a little poorer, stealing their data from them, little by little, invisibly. This trend is not your friend.

How the One Stop Shop Flop undermined Europe’s grand privacy law

2025-12-04T12:28:00.000+01:00

I believe in privacy, and I believe in the need for privacy laws to guarantee privacy. I’ve been calling for serious global privacy laws for two decades:

http://news.bbc.co.uk/2/hi/technology/6994776.stm

Europe has been the continent with the world’s most serious and strict privacy law, but only on paper. It’s been almost a decade since Europe’s landmark privacy law, the GDPR, was passed. It’s fair now to assess its impact. The European Commission is taking a look at certain revisions, mostly irrelevant and mostly in the direction of loosening some of its paperwork provisions. Meanwhile, leading privacy advocates are already warning against weakening the GDPR: https://www.theguardian.com/commentisfree/2025/nov/12/eu-gdpr-data-law-us-tech-giants-digital

I have a lot of experience working with these laws, and I have a clear opinion: the GDPR has failed. It has not met any of its key goals. If you have any doubt, ask yourself the simple question: do I have more or less privacy today than a decade ago? There are many reasons for this failure, but number one is the One Stop Shop Flop.

The European lawmakers passed a strict privacy law, on paper, with massive potential fines (up to 4% of worldwide turnover) for non-compliance, on paper. But then, in a massive blunder, the European lawmakers created a one-stop shop notion, meaning that non-European companies, like Chinese and US Big Tech, could pick any EU country to regulate them. Guess what, they picked Ireland, their longstanding tax haven, as their regulatory haven too. Before this law, all 28 European countries could enforce privacy laws against Big Tech. After this law, only Ireland. Hallelujah for American and Chinese Big Tech…

But let’s keep things simple. A law with no enforcement will not be respected. A law that is “enforced” by a small regulator, based in a small country, with a small staff, in a country that makes its money by being a tax haven to American and Chinese big tech…was tasked with enforcing this law on behalf of 450 million citizens of the EU?... I’m hardly the only person to criticize this farce: https://noyb.eu/en/former-meta-lobbyist-named-dpc-commissioner-meta-now-officially-regulates-itself

I love to take early morning walks in a park near my home. Every morning I see the same spectacle: a few adorable dogs chase the local squirrels. The dogs bark and wag their tails, the squirrels scamper and scurry, and the humans chuckle. It’s fun for all, because everyone knows…the dogs will never catch a squirrel. And indeed, no one in the “lead” privacy regulator for Europe, in Ireland, has ever caught a Big Tech squirrel, and never will. The Irish regulator has never imposed the fines that the European lawmakers envisaged.

Meanwhile, the European Commission proposes to fix this iconic privacy law…not by asking why this key law has failed, but by suggesting its paperwork documentation obligations could be streamlined. Meanwhile, the one stop shop flop continues, and the Big Tech squirrels are not worried about any privacy law enforcement. The losers, of course, are the 450 million Europeans who were promised a strict privacy law.

The lesson for the future, in particular for AI, is clear. You can pass laws (as Europe already has done in its AI Act), but a law with no enforcement won’t be respected. Ask the squirrels.

Data centers coming to a town near you

2025-11-12T21:43:00.000+01:00

You may think of your personal data as a digital asset, but it does have a physical home. It exists in a data center (or multiple data centers) at rest. As you’ve read in the press, tech companies are building data centers at a frenzied pace. The massive computing needs of AI require them to build them as fast as they can, regardless of the cost. It’s now becoming debatable whether we’re living through a period of boom or bubble in data center construction: https://www.theguardian.com/technology/2025/nov/02/global-datacentre-boom-investment-debt

Google plans to spend something like 90 billion dollars investments, mostly on data centers (not humans) this year. This week alone it announced something like 5 billion to be spent on data centers in Germany. This made the German politicians very happy, crowing about Germany’s high-tech investment environment. https://www.tagesschau.de/wirtschaft/unternehmen/google-investition-deutschland-klingbeil-100.html

But should local people really be happy to see a data center built in their backyard? Should politicians really welcome them? Benefits are few: data centers create very few long-term jobs, mostly junior technicians. The servers, chips and high value tech work are done remotely, often on the other side of the planet. Harms locally can be more significant: data centers consume vast amounts of power and water (to run and cool them). This can put stress on either the electricity grid or the water supplies or both, and sometimes leads to rising electricity costs for all. https://fortune.com/2025/11/08/voter-fury-ai-bubble-high-electricity-prices-offseason-elections-openai/

The environmental impact of data centers is now the subject of study. https://www.theguardian.com/technology/2025/nov/10/data-centers-latin-america

Back in 2007, I was asked by my then-employer Google to officiate publicly the opening of its big new data center in Belgium. The investment was valued at around 250 million. (Compare that to Facebook’s recent announcement of planning to spend 600 billion on data centers …https://finance.yahoo.com/news/meta-plans-600-billion-us-175901659.html ). But back then, 250 million was a big deal. Even the Belgian Minister President (later Prime Minister) Elio Di Rupo came along to celebrate it, and he and I took a long pleasant stroll through the town of Mons after the ceremony. Like the German politicians this week, welcoming Google’s data center announcement.

But shouldn’t we all know better now? Data centers are the physical concrete of the industrial infrastructure of the digital world. The industrial revolution needed coal mines and steel mills, until it didn’t. The Google data center I opened in Belgium was in a depressed region full of abandoned coal mines and steel mills. I fear history repeats itself.

I understand that local politicians will welcome “investment” in their countries, even if 99% of the value of that investment comes from servers and chips and software created far away. But we all need to look at the local environmental cost of these data centers. If not, you’re just breathing the pollution of someone else’s beautiful shiny Ferrari.

Taking the Humans out of Human Resources

2025-11-09T17:34:00.000+01:00

We all know the trend: AI is replacing human workers. We all know that trend is accelerating. One of the industries where AI is replacing humans is...ironically, human resources. And even more ironically, increasingly, we humans will be hired, or not hired, as decided by an AI agent. I think we deserve to know why. Privacy law requires it.

Let’s take the example of a company called Workday. It’s a human resources outsourcing company using AI to replace humans in human resources. https://www.workday.com/en-us/homepage.html

Lots of companies, including my former employer Google, have outsourced human resources functions to Workday. When I worked at Google, I was forced (like all employees) to provide very detailed, sensitive information to Workday. I had little idea what Workday did with that information, or where they sent it (e.g., worldwide back to the US, I’m guessing, since it’s a US-based company?). Workday has been sued in the US, alleging Workday’s tech discriminated against people over 40 from getting hired. https://www.cnn.com/2025/05/22/tech/workday-ai-hiring-discrimination-lawsuit

Companies like Workday are now using AI to replace humans in the human resources functions, like filtering CVs or conducting “interviews” between AI agents and human job applicants. You will be hired, or not, based on the automated decision made by an AI agent.

AI agents work by defining “success” based on the models on which it was trained. Then they assess job candidates or job workers based on how closely they correlate to these models of “success”. They can collect and assess hundreds or thousands of characteristics, far more than any human could. I’m concerned that AI models will reinforce whatever forms of bias exist in the training models of “success”. Imagine an AI “interview”: what data is it collecting? It’s not collecting data like a human interviewer. An AI “interviewer” can measure pupil dilation, eye movements, head angles, speech rhythms, vocabulary spectrum, verbal biometrics…you get the idea, collecting vastly more sensitive data than a human could in order to build an assessment.

Let’s take the example of an AI agent that is tasked with hiring a CEO for a tech start-up. The AI agent will search the public data of the world about successful CEOs of tech companies. Which characteristics will it find? How will it assign a weight to each of those characteristics? Let’s play a thought experiment, based on fairly obvious assumptions of what an AI agent would find based on the current crop of successful tech CEOs. I’m not going to name names, but they’re all household names anyway.

Ethnicity: high correlation for being Jewish or Indian. Negative for being Hispanic or Black

Gender: high correlation for male. Negative for being female.

LGBT: positive correlation for being cis male gay. Negative for all other categories of LGBT.

Neurodivergence: high correlation for being on the “high functioning” autism spectrum. Negative correlation for all other categories of neurodivergence.

Personality: high correlation for narcissistic personality disorder. Negative correlation for introversion.

Age: high correlation for 25-40. Negative correlation for over 40, increasing by age.

You may want to dispute some of my assumptions above. I’ll accept that. The list of characteristics above may vary from function to function. What’s “good” in the list above for a tech CEO may be “bad” for a mid-level tech worker.

As companies turn over their human resources functions to outsourcing companies that have in turn outsourced these functions to machines…it’s essential that we ask these companies the criteria they’re using to make these automated decisions about our most basic human aspirations, like getting or keeping a job.

I know privacy regulators are under-resourced. So, to my friends in the privacy regulatory community, I’ve drafted the simple questions you should send to companies in your countries.

Do you engage third party companies to provide human resources functions? If so, please provide their names.

Do you provide these companies with criteria to apply to recruiting or evaluating your employees? If so, please list them. Can your employees or applicants object to sharing their data with these companies?

Do these companies transfer such data outside of your country, e.g., to the US? If so, under what legal basis? Can your employees object to this?

Do these companies use AI? If so, on what data has it been trained?

What steps does it take to eliminate bias (relating to race, age, gender) in its automated decisions?

What transparency do these companies provide in terms of how it is making its automated decisions? List the criteria used to make these automated decisions. Can an employee or applicant object to the use of such automated decisions? What would be the consequences of such an objection? Is there an ability to appeal such automated decisions? Is there a human in the loop?

Look. I know we (humans) are on a path to automating more and more of our human jobs. Human resources is one of them. Technology is taking humans out of human resources. But it’s up to us to keep humanity in the machines. Privacy law is central to this. Let me conclude by citing Google’s own AI-generated overview of the topic:

European privacy laws, primarily the EU's General Data Protection Regulation (GDPR) and the upcoming AI Act, regulate automated decision-making by providing individuals with the right to an explanation, to human intervention, and a qualified right to opt-out of certain automated decisions. The GDPR's Article 22 prohibits decisions based solely on automated processing that produce legal or similarly significant effects, unless specific conditions are met, such as explicit consent or a necessary contract with safeguards. Key rights include transparency, the right to a human review, and the right to challenge such decisions.

Worried about AI? Don’t worry, it will cure cancer

2025-10-30T12:09:00.000+01:00

You have surely read the dire predictions of the leaders of the field of AI: everyone from the leaders of OpenAI, Anthropics, Bill Gates, etc etc have been speaking about the massive challenges that AI will thrust onto our societies, including in particular the massive imminent wave of job destruction. So, it’s refreshing to hear from Google’s President that AI “will cure cancer”. https://fortune.com/2025/10/26/google-ruth-porat-cure-cancer-in-our-lifetime-with-ai/

Some of you will be cynical, and suggest that she is just whitewashing AI in the PR interest of her own employer, and her own. She’s repeated that same line about AI curing cancer more times than Brittney Spears has gotten wasted.

I’d like to hear more leaders of companies building AI tools engage in a public discussion about the good and bad consequences of their inventions. The tech industry is famous for privatizing gains and socializing losses. In other words, building their businesses and their share prices based on the good use cases of their inventions, but letting other people, governments or societies deal with the negative fall-out. Heads, I win, Tails, you lose. For example, a company could make a fortune “curing cancer”, but would it be held responsible if the same AI tool that it built to cure cancer could also be re-purposed to build bio-weapons?

Let’s take the example of using AI to screen CVs. Many people looking for jobs today will be auto-rejected by an AI bot. There’s no transparency, and they won’t be told why. One reason might be that they’re screened out for being “old”, amongst many other possible, but non-transparent, reasons. https://www.wired.com/story/ageism-haunts-tech-workers-layoffs-race-to-get-hired/ And indeed, in the US, a lawsuit on precisely this topic has been launched, accusing Workday of using AI tools to discriminate against older job applicants: https://www.hklaw.com/en/insights/publications/2025/05/federal-court-allows-collective-action-lawsuit-over-alleged

AI in recruitment is pretty simple. Ask the AI to study the characteristics of “successful” employees, and go find me more like them. So, if the data about “successful” employees skews heavily to the age range of 25-39, well, the AI will look for more of the same and auto-reject the rest. That’s how AI will reinforce discrimination in our societies. The category of ageism is not unique: sexism, homophobia, racism (some races, but not others) are all categories that AI will discriminate against, simply to pursue the goal of advancing the types of people who meet its training model of “success”.

Europe’s privacy law (GDPR Article 22) has a very clear provision dealing with machines that make automated decisions:

The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her.

I’m not aware of any serious regulatory attempt to enforce GDPR Article 22 against companies using AI to screen job applicants. But they should. If companies are using machines to decide whether to hire or auto-reject an applicant, then those decisions “significantly affect him or her” under the law.

AI has already begun to accelerate its trend to replace human workers with machines. Companies are already replacing entry-level functions with AI agents, making it hard for young workers to get their first rung on the employment ladder. Older workers have long been pushed out of Silicon Valley. The trend is just starting, but even today, hardly a day goes by without some big company announcing plans to slash human jobs and hire machines. https://seekingalpha.com/news/4508911-amazon-plans-to-cut-up-to-30000-jobs---reuters

You may despair of getting or keeping a job in this new environment. But you don’t have to accept, in Europe at least, that a machine will make an automated decision not to hire you, in violation of existing law. Often regulators don’t act, until complaints are filed. If your CV is being auto-rejected by a machine, you can file a complaint with your local data protection authority, which might prompt them to intervene.

AI is a tool for automated decision making: that’s the whole point of it. I cited one example from the world of job applications, but there are thousands of other examples, today and soon. It’s high time to start applying and enforcing the laws against automated decision making. The biggest imminent disruption in human history is around the corner: so far, I can confidently state that privacy regulators have had very close to zero impact on the development of how AI is being developed and used. The Italian Garante is one of the few to have tried, and I applaud them for their leadership. The others seem to be content with conducting blah blah public consultations. If we want AI to respect human values and the laws, we need to speed up, urgently, because the machines won’t slow down. Enforcing the laws on automated decision making would be a good place to start.

The Age of Discovery, seen from Seville

2025-10-23T16:16:00.000+01:00

I enjoyed a week in enchanting, intriguing Seville. The photo is the tomb of Christopher Columbus in the Seville Cathedral. In the Age of Discovery, Seville had a Spanish monopoly on ships to/from the New World. Historic Seville can teach us a lot about our own AI-driven age of discovery. Both eras have a lot in common, driven by science, greed and missionization.

Europe won that technological race: compared to the indigenous populations of the “New World”, it had superior sailing/navigation/mapping tech, it had superior military tech, it had deep capital markets to fund the expeditions, and it had a belief in its cultural and religious superiority. That’s a good description of the people leading the AI race today.

Europeans in the Age of Discovery expanded human knowledge and science dramatically, and AI will do the same now. But even though some actors were driven by science and a pure search for knowledge, most were driven by greed. Leaders in the field of AI are now accumulating vast (perhaps bubble) riches, just as the riches of the New World poured into Seville in the Age of Discovery. As a tourist in Seville, you can still visit the architectural gems financed by plundering the New World indigenous populations. Then as now, some people got very rich, but most people got much poorer. The Spanish royal house got rich, the indigenous populations were plundered. The tech bros of today have gotten obscenely rich, the soon-to-be-unemployed legions of workers of today replaced by AI agents will get poor.

The biggest losers of the Age of Discovery were the indigenous populations, wiped out by European-introduced diseases. 90% of indigenous populations were wiped out within one century of contact with European colonizers. AI will probably do the same to us: it’s becoming a consensus that superintelligence (whenever that happens) will eventually similarly cull or eliminate homo sapiens. Lots of leaders are calling for a (temporary) ban on developing superintelligence, until our species can figure out how to build this safely. My former colleague, Geoffrey Hinton, a Nobel prize-winning AI genius is amongst them.

History tends to repeat itself. As we enter into our own new AI-driven age of discovery, ask yourself if you think you and your society will become winners or losers. A lot of people today think they’ll be winners: tech bros (obviously), governments and businesses looking for new tech-driven growth and profits, scientists, entire countries like China or the US which are currently leading the race. But lots of people will be losers: in particular, looming job destruction and unemployment, in turn leading to social disruption, which in turn historically tends to lead to revolutions. Which do you think you’ll be, winner or loser? Even if AI doesn’t destroy humanity, yet, it may well destroy democracy. It will destroy privacy too (I’ll blog about that separately).

Privacy is anchored in the idea of the dignity of the individual human being. There wasn’t much dignity in being an indigenous person dying of smallpox during the Age of Discovery, or an African victim of the trade routes of the Age of Discovery that evolved into the slave trade. Can we do better today? Machines don’t believe in privacy: they consume data to output data to accomplish a task. The rise of AI is the challenge of our age. You might ask where to start: how about stopping private companies from plundering other people’s intellectual property or personal data to train their AI models, as the Spanish conquistadors plundered the wealth of the indigenous populations.

Lots of us need to step up to confront this challenge. Or we can leave it in the hands of the tech bros and gullible politicians and impotent regulators, who are welcoming AI like Montezuma welcoming the Spanish.

The world’s largest surveillance system…hiding in plain sight

2025-10-01T15:15:00.000+01:00

The world’s largest surveillance system is watching you.

It’s capturing (almost) everything you do on the web on (almost) every website. And it’s hiding in plain sight.

And it’s “legal” because it’s claiming that you know about it,

and that you consented to it.

But do you know what it is?

Do you know what “analytics” is? Websites use analytics services to give them insights into how their users interact with their sites. Every website wants to know that. And analytics providers can give them that information. For example, an analytics provider can give detailed statistical reports to a website about their users and how they interact with its site: how many people visited the site, where did they come from, what did they view or click on, how did they navigate the site, when did they leave/return, and many, many other characteristics. This data can be collected and collated over years, over thousands or millions of users.

There are many providers of analytics service, but according to analysts, there is only one 800-pound gorilla, Google Analytics.

“Google Analytics has market share of 89.31% in analytics market. Google Analytics competes with 315 competitor tools in analytics category.

The top alternatives for Google Analytics analytics tool are Tableau Software with 1.17%, Vidyard with 0.78%, Mixpanel with 0.59% market share.”

And according to other third party analysts: “As of 2025, Google Analytics is used by 55.49% of all websites globally. This translates to approximately 37.9 million websites using Google Analytics.”

You get the point: one company, one service is capturing the bulk of the web traffic on the planet. Websites get statistical reports on the user interactions on their sites. Google gets individual-level information on the actions of most everyone on the web, on most websites, click-by-click, globally. Wow.

Legally, a website that uses Google Analytics is contractually obligated to obtain "consent" from its visitors to apply Google Analytics. But often the disclosure on those websites is cursory, or even incomprehensible: “we use analytics”, or “we use analytics software for statistical purposes”...which sounds harmless, but hardly would explain what’s happening to the average user. Technically, what happens is simple, but invisible to the average user: when they click on a website, that website auto-transfers to Google, in real time, detailed information about every step a user takes on its site. What’s happening is very simple. A site using Google Analytics incorporates a small piece of code on its site which auto-transfers to Google, in real time, information about every interaction its users have: every visit, every click, and information about each of those visitors, on an identifiable basis.

In fairness, Google Analytics has some privacy protections. Its reports to its client websites are statistical, rather than reports at individual users. But even if the websites don’t get information about users at an individually-identifiable level, Google does…. And Google does not do cross-site correlation, i.e., it does not profile users across sites, for Analytics purposes. (Note, Google does exactly this cross-site correlation in the context of its Ads businesses, but that’s a different topic than this blog.)

All this is “legal” if it’s based on consent. A phrase disclosed in a privacy policy, or a cookie notice, no doubt you’ve seen, or maybe clicked on, is deemed to constitute “consent”. But really, did you or the average user have a clue?

I’m in the school of believing that analytics tools represent a relatively low level of privacy risk to individual users. But what do you think if one company is getting real-time information about how most of humanity is engaging with websites on a planetary level? A user goes to any random site, but their data also auto-transfers to Google, did they know? Since the scale of this service vastly exceeds any other service on the web, the scale of this data collection is the largest on the web. Please respond with a comment if you can think of anything of similar surveillance scale. I know you can’t, but let’s engage in the thought experiment. I’m not picking on Google (I love my former employer), but in this field, which is essential to privacy, it’s the 800-pound gorilla, surrounded by a few mice.

And the photo, if you’re interested, is Chartres Cathedral, built in the era when we believed only God was all-knowing.

The Irish Backdoor

2025-09-24T11:35:00.000+01:00

It’s not a gay bar in Dublin, the Irish Backdoor, sorry if that’s why you clicked on this blog. It’s how non-EU companies, like tech companies from the US and China, use the “one stop shop” mechanism to evade the privacy regulations of 26 countries to be regulated instead by the Irish regulator, the gentle golden retriever of privacy enforcement.

I am expanding on my blogpost below. But now I’m revealing something new. How most of the non-EU companies, like tech companies from the US and China, have no legal right to assert a claim to be regulated by the one stop shop. Fiction or fraud? Let me explain.

Legally, a non-EU company can only claim the benefits of the one stop shop if the decisions regarding data processing in Europe are made there.

Let me suggest a reality test. Most companies from outside the EU claim to benefit from the one stop shop in Ireland if they do the following: 1) create a corporate entity in Ireland, 2) write a privacy policy (or ask ChatGPT to write one) that tells users that the Irish corporate entity is the “controller” of their data in Europe, and 3) has some minimal presence in Ireland, like appointing some employee as a “data protection officer” for the entity. All this can be done in a day, and with a tiny local Irish staff. But…does this meet the legal test?… that the data processing operations in Europe are being decided by this Irish entity?

Most tech companies build products in their homes, Silicon Valley, China, etc. They then roll out these products globally. Usually these products are identical worldwide, except for language interface translations. In those cases, does anyone really believe that their Irish subsidiaries are really the decision-makers for how the data will be processed for their services for their (millions) of European users? Perhaps that is the case for a few large non-EU companies with large operations in Ireland. For all the others, it’s hard to believe.

Maybe it’s an innocent fiction for a company from China or the US to claim it is “established” in Ireland to evade the privacy laws of 26 EU countries with millions of users. Or maybe it’s a fraud…?

(Final note, as a former employee of Google, I must point out that nothing in this blogpost is meant to suggest anything regarding that particular company. Google has a huge workforce in Ireland).

Meanwhile, non-EU companies are getting an easy ride in Europe, while their EU company competitors aren’t. I just don’t think that’s fair to EU companies or to EU users.

Why does every US and Chinese company want to go to Ireland?

2025-09-22T10:55:00.000+01:00

Ireland is one of the biggest winners of the EU 27 construct. It has established itself as a tax and regulatory haven for foreign (non-EU) companies. Virtually all Chinese and American companies, in particular in tech, rush to “establish” themselves in Ireland. In exchange, they get to pay a low corporate tax rate (even if their users and their money is made in the other 26 EU countries) and they get to benefit from the light-touch privacy regulation of Ireland.

You’ll recall that Europe’s tough (on paper) General Data Protection Regulation of 2018 created the concept of a one-stop shop for foreign companies. So, any Chinese or American company could pick one of the EU countries as its “establishment”. Of course, they all picked Ireland, given its universal reputation for light-touch tax and regulation. It is entirely a different debate about why/how Europe made this blunder: in effect, it gave a massive advantage to foreign companies over domestic European companies. A French/Italian/Spanish company would be regulated by their domestic French/Italian/Spanish regulators, who take privacy seriously, and would sanction non-compliance. But a Chinese or American tech company would do business in all those countries, while benefiting from the Irish regulatory culture, as gentle as an Irish mist.

Occasionally, a European regulator would try to take on an American or Chinese company in the field of privacy. https://www.cnil.fr/en/cookies-placed-without-consent-shein-fined-150-million-euros-cnil

But this action wasn’t based on the core European privacy law, the GDPR, but on a rather obscure law about other things.

The Trump administration has defended American companies in Europe against what it claims are discriminatory regulatory actions. https://www.lemonde.fr/en/international/article/2025/09/06/eu-commission-reluctantly-fines-google-nearly-3-billion-despite-trump-threat_6745092_4.html# It was therefore not a surprise to see the French regulator announce fines at the same time against one American and one Chinese company. But it is surprising to see the Trump administration rushing to defend one of the most Democratic-leaning companies in the US.

Indeed, Europe does discriminate, in the field of privacy, in favor of non-EU Chinese and American companies, due to the one-stop-shop Irish backdoor. One can only assume European dysfunctional politics led to this absurd result, from a European perspective. Hundreds of millions Europeans depend on a small Irish privacy regulator to ensure that the gigantic American and Chinese tech companies respect European privacy laws. Hilarious.

All of this might seem like trivial corporate politics, but the consensus is growing that humanity is allowing the tech industry to put us (I mean, our entire home sapiens species) on a path to doom. https://www.theguardian.com/books/2025/sep/22/if-anyone-builds-it-everyone-dies-review-how-ai-could-kill-us-all Even if we’re doomed, can we at least put up a fight?

Hi Privacy Pro's: where's your mojo?

2025-08-28T11:38:00.000+01:00

I’ve been committed to the field of privacy for 3 decades, and I’ve had the pleasure to mentor multiple generations of smart and committed people to the field. But I can’t remember a time when the profession felt more disempowered and disrespected than now.

Where have all the senior privacy leaders at Big Tech gone? Virtually all the Big Tech companies have lost (or fired) their most senior privacy leaders this year. The most senior privacy leaders at Microsoft, Google, Facebook, and Apple have all exited this year, or recently. These are the companies that process vast amounts of personal data, so it’s not a minor question to ask why they’ve lost their most senior privacy leaders. Undoubtedly, each person who exited their employer will have their own story, and I won’t tell it, even if I know it. But if these companies have lost their most senior privacy leaders, who is left there to ensure that these companies respect their users’ privacy?

The privacy leaders of my generation (and I knew them all) all shared one characteristic: they advocated for good privacy in their organizations internally, and they worked collaboratively with regulators to find solutions when required. But perhaps the collaborative model is no longer the fashion in Silicon Valley: perhaps the truculent, cage-fighting ethos has the ascendancy, reflecting the personalities of some of its leaders: media-hungry, kick-boxing, “I am Caesar” and anyway, I have a survivalist bunker in case it doesn’t work out. In that world, you don’t want privacy leaders, you want privacy litigators. Privacy litigators can make an easy meal of the average privacy regulator, who have tiny technical and litigation resources.

Privacy only makes sense as a human value, since its only purpose is to protect the autonomy and dignity of an individual human being. In an age when Big Tech fires many thousands of workers (in the name of “efficiency”), often without warning, by email at 2 am, with immediate effect (I don’t need to name names, do I?), it’s fair to ask what respect they have for individual human beings. If you don’t respect your own employees as human beings, why would you respect your users, or their or anyone’s privacy?

Try to read a privacy policy, when you randomly click on some website. It will inevitably begin with the phrase: “We care about your privacy”. Then it will go on to list the innumerable ways that they plan to violate your privacy, to track and profile your data, and to share it with hundreds of their “partners”. You cannot possibly understand these privacy statements, and neither can I. They’re not designed to explain privacy practices: they’re designed to create a veneer (or fiction) that their companies’ data collection practices have been disclosed, and that users have somehow “consented” to them. Of course, you can’t consent to something that you can’t understand, but a click looks like consent, so that’s all these companies are seeking. The latest atrocity is the attempt by sites to ask you to consent to tracking your “precise location”. Usually this phrase is buried innocuously deep inside the privacy statement. If you are dumb, or bored enough, to click “I accept”, these companies will track your precise location (within meters) every time they encounter you on the web, and share that with their hundreds of partners, and store your precise locations forever, and heaven knows what they’ll do with that. Nothing creepy there?

It’s all about (sharing) the data, stupid: privacy meets antitrust

2025-06-12T11:59:00.000+01:00

I spent time with a group of privacy experts recently. We were discussing the intersection of privacy and antitrust law. Traditionally, these two fields were very separate, with separate laws, separate regulators, and separate practitioners. But the rise of the data-processing monopolies like Google and Facebook is forcing these two fields to converge. When a monopoly like Google Search or Facebook is based on processing vast amounts of personal data, and when no competitor could possibly compete with these data-gorged monopolies, well, it’s obvious that antitrust law should consider forcing these monopolies to share data with potential competitors. Otherwise, these monopolies will carry on with their “data barrier to entry”. Data is an essential input into any of these existing or future services.

Existing monopolies, like Google Search, do not want to share their data with potential competitors. Duh. So, they are making public arguments that such sharing would create a serious risk of violating the privacy of their users. But is that true?

Google has resorted to public blogging to warn its (3 billion) users of the risks of court-mandated data sharing. “DOJ’s proposal would force Google to share your most sensitive and private search queries with companies you may never have heard of, jeopardizing your privacy and security. Your private information would be exposed, without your permission, to companies that lack Google’s world-class security protections, where it could be exploited by bad actors.” https://blog.google/outreach-initiatives/public-policy/doj-search-remedies-apr-2025/

Now, let’s unpack that statement. Google is clearly stating that it collects “your most sensitive and private search queries”. Its privacy policy makes it clear that it collects, retains and analyzes that data to run and improve its own services (not just Google Search). So, Google clearly analyzes your “most sensitive and private” data, despite the privacy issues to you: the privacy issues, according to Google, only arise if that data is shared with other parties.

Now think about Google’s money machine, its ads network. Doesn’t that network do exactly what Google is here claiming is a terrible thing for users’ privacy? Google ads network collects vast amounts of its users “sensitive and private” surfing history, and shares it with “companies you may never have heard of”. Indeed, that’s exactly what the ads network does today. Not coincidentally, an unrelated antitrust monopoly case is underway regarding the Google ads monopoly. So, let’s be clear, in the context of Google Search, Google claims sharing data with third parties would be terrible for users’ privacy, but in the context of Google ads network, all that sharing is just fine…

Privacy professionals should take a clearer look at the privacy implications of any court ordering Google to share Search data with competitors. Would that really raise any privacy issues? Some experts in the field are starting to discuss the issue: https://www.hklaw.com/en/insights/publications/2025/04/google-search-data-sharing-as-a-risk-or-remedy

Search is based on data mountains. They are different mountains. Each category of the data mountains has different privacy implications. We need to unpack data-sharing into its different categories to assess whether it has any impact on privacy.

The Index: the biggest data mountain is the Search index. That’s the index that Google Search creates by crawling the entire public web. It’s one of the largest, if not the largest, database on the planet. But it’s not a privacy issue: it’s just crawling the public web. Of course, there is public data on the public web, but it’s not a privacy issue to force Google to share such data with other parties, who could also access it on the public web.

User interaction data: with its 3 billion users, and over 20 years of operation, Google Search has the largest database of user interaction data on the planet. I’m guessing it’s 1 million times larger than its nearest competitor Bing. (Google can correct my guess if it wishes to.) This user interaction data is essential to teach a Search engine’s algorithm how to guess what someone intends to find when they type a query. If you have billions of examples of what people are searching for, you can train your search algorithms accordingly. If you don’t have that data, you don’t have a chance. So, would it be a privacy issue, as Google menacingly suggests in its blog post, if it were forced to share such data? It depends: yes, if it were forced to share search histories (i.e., search logs) with all of the personally-identifiable data that Google collects and shares. No, if it were forced to share anonymized data sets, such as anonymized search logs.

Fortunately, many years ago, Google introduced its policy to anonymize search query logs, after a number of months, in the interests of users’ privacy, and to respond to regulators’ pressure. I know something about that, since I worked on that privacy initiative, with my great former colleagues.

https://publicpolicy.googleblog.com/2008/09/another-step-to-protect-user-privacy.html

There is no privacy issue, none at all, with forcing a company to share anonymized user interaction data.

I get that Google is blogging as part of its anti-antitrust litigation strategy. It really, really, doesn’t want to share its data with potential competitors. Litigators will advance their clients’ interest, as best they can. The rest of us 3 billion users of Google Search can assess the intellectual honesty of their arguments. As far as I am concerned, there are profound privacy issues on the web: forcing the Google Search monopoly to share its non-personally identifiable data with potential competitors is not a privacy issue.

Debating Privacy in Venice

2025-04-29T10:18:00.000+01:00

I’m looking forward to seeing lots of old friends at the upcoming Venice https://privacysymposium.org/

For many years, I attended and spoke at privacy conferences around the world.

I believe in sustaining a dialogue amongst privacy professionals, regulators, academics and advocates.

I always learned a lot from these events, and I did my best to contribute to the debates as well.

I also believe in building human connections to the people in this field, and I’m happy to count many of them as personal friends.

This year, I’ll join a distinguished group of regulators and practitioners on a panel entitled: Privacy and Antitrust.

This should be interesting! I have, after all, spent 30 years guiding Microsoft and Google…

Like privacy itself, Venice is precious and fragile.

We’re lucky to be there together in May, before another tech monopolist rents the city for himself in June.

A Gaggle of Monopolies

2025-04-23T12:04:00.000+01:00

One of the peculiarities of monopolies in the age of Big Tech is how they tend to leverage quickly into a group of monopolies.

Historically, building a monopoly was a rare business event, and it usually just happened in a single industry, like oil or finance.

But the tech industry is different: businesses build one monopoly (legally, let’s assume).

Then they quickly manage to leverage it into multiple monopolies across an array of businesses.

You can read recent press reports about antitrust enforcement actions against Google and Meta, to take those prominent examples.

Traditional antitrust/competition law developed to restrain individual monopolies from leveraging their existing monopolies unfairly into new markets. But historical law seems to struggle with how to address this new phenomenon of companies that develop a portfolio of monopolies. Of course, these tech companies leverage their monopolies to develop and support each other, in particular by sharing user data, given that all these monopolies are based on processing vast amounts of user data. The more you have, the better you can leverage into a new market. That’s why this antitrust/competition conundrum is also a privacy challenge. Monopolies that process personal data and share them across their portfolio of services are processing personal data at a scale unprecedented in human history. Europe took a first step to try to address this problem with its Digital Markets Act.

We don’t have a legal word for a portfolio of monopolies. Calling a company a “monopolist” doesn’t capture the nature of a portfolio of interlocking monopolies. So, I looked to the wildly colorful words of English vocabularies to describe a group of animals, for inspiration.

A bloat of hippopotamuses

A parliament of owls

A gaggle of geese

A flamboyance of flamingos

A murder of crows

A company of parrots

A charm of finches

A shiver of sharks

An aggregation of snakes

A gamble of alligators

A skulk of foxes

Antitrust/competition law will have to come up with new tools to deal with this new phenomenon of portfolios of monopolies, as will the field of privacy. Any remedies that the authorities impose will need to take into account the nature of these interlocking monopolies. And yes, forcing a company with a portfolio of monopolies to divest one of its monopolies might be the right way forward, or to stop it from acquiring new ones. I doubt though that a “murder of crows” will suffer terribly if it loses one crow.

But first, let’s find a name: a “gaggle of monopolies”?, a “bloat of monopolies”? Any of the above might do, with all due respect to the animals. I’m happy to let Llama or Gemini choose.

The end of US clouds in Europe?

2025-04-07T10:27:00.000+01:00

How should Europe riposte to the Trump tariff wars? I visited with some friends in Madrid recently.

I was surprised how quickly the consensus was reached: “it’s time to liberate ourselves from the US cloud providers.”

As background, if you don’t already know, the global market is dominated by three US cloud service providers: Amazon, Microsoft, and Google, in that order of dominance. It’s no secret that all three of these entities run giant tech monopolies in other markets, which they leveraged into the cloud market. The cloud market isn’t particularly high-tech. It’s more about scale: the bigger you are, the more you can build a global infrastructure network and lower costs for all. There are lots of smaller local competitors, but none can match the scale of the US giants.

But Trump has changed the global understanding of the risks of entrusting your critical national infrastructure to three US companies. Could Trump order these US companies to terminate their services, immediately, in Europe or any other country? What was once unthinkable, is now a possible (hopefully, improbable) scenario. European business and political leaders are now asking what they would do if the US government ordered these US cloud services to terminate them. Bedlam hardly describes what would happen. Relying heavily on US cloud providers is reckless, leaders are realizing.

There have long been criticisms of US cloud providers. In Europe, for example, they have been criticized for utilizing tax haven structures in Ireland and Luxembourg to earn vast revenues, while paying tiny taxes. Want to compare the income tax rates of a Madrid bus driver to the tax rates paid by Amazon, Microsoft, Google in Europe? Guess which is higher.

Europe already has lots of laws on the books that could be used to drive forward to a US cloud-free future. Europe’s main privacy law, the GDPR, has prohibitions on transferring personal data from Europe to the US, as long as the US does not have “essentially equivalent” data protection laws. (I’ll blog on this separately, but the chances for the European Court of Justice to decide that any new data transfer scheme between Europe and the US meets that test are about as high as my chances to win the Paris marathon next weekend.)

In purely trade terms, the US runs a huge trade surplus with Europe in digital services, with cloud services high on the list. As Europeans look at ways to riposte to Trump’s trade war, consider the risks to US cloud providers. If I was a US West Coast cloud executive, I’d be quaking in my birkenstocks. Meanwhile, the sky over Madrid is cloud free.

Mirror, Mirror, on the wall, who’s the wokest of them all?

2025-04-03T11:18:00.000+01:00

I spent most of my life in woke organizations, namely Harvard and Google. Both of those institutions have lots in common: both have very talented and intelligent communities, and both are amongst the wokest organizations on the planet. They were both meritocracies, or at least, they used to be, and to a lesser extent, still are. But both became radical converts to wokism, and both developed cancel-cultures to stifle any dissenting voices.

Harvard appointed a black female President, with an underwhelming academic record, who later resigned, as you’ll recall, in part because she couldn’t figure out if calling for the genocide of Jews was against Harvard policy. I was in the Google legal department, and so I celebrated when Google also appointed a black female general counsel, as my department’s boss. I’m sure she’s brilliant, and like me, a former Harvard student.

The one thing that the woke brigades all agree on is simple: the insufferable inherited privilege of the white male. I wasn’t surprised to see a former Googler sue the company for discrimination: https://nypost.com/2025/04/02/business/google-executive-discriminated-against-male-employees-bombshell-lawsuit-alleges/

To be clear, I know none of the people mentioned in the article above, and I have no opinion on the merits. Even if I did know them, I wouldn’t publicly comment on it, out of loyalty to my former employer.

The truth, however, is that it’s hard to be a white male at Google. And even harder to be an older white male at Google. When I left Google, I was a 60-year-old white male. I never met another one at Google, not a single one. You might be surprised that there was even one.

23andMe, Privacy zombie

2025-03-28T13:38:00.000+01:00

Will it finally go away? 23andMe is filing for bankruptcy. https://www.npr.org/2025/03/24/nx-s1-5338622/23andme-bankruptcy-genetic-data-privacy

23andMe's entire business model peddled pseudoscience for years. https://www.theguardian.com/commentisfree/2025/mar/27/geneticist-mourn-23andme-useless-health-information

However absurd the slick test results were presented to 23andMe customers, however absurd its “insights” into genetic health risks or ancestry, it did work wonders at identifying one genetic trait: stupidity. If you spit into a test tube, and sent your saliva to this shadowy company with a long history of privacy breaches, then I can conclusively determine that you are genetically…stupid. Your genome is your most personal, sensitive, unchanging identifier. You handed over this data to a company built on data mining for cheap fun, and even paid them money for the privilege. They will retain this data forever, unless you take steps to delete it, assuming that they actually delete it when you make that request, which is a fair question given their shadowy history of privacy practices in the past. At least, please go try to delete it and hope it actually is deleted. Whether they actually delete it or not, you’ll never know.

Now, your genetic data is considered this bankrupt company’s “asset” that they plan to sell to the highest bidder. Your genome has become their bankruptcy asset. You might have trusted 23andMe with your genetic data, but will you trust whoever buys it in a bankruptcy sale?

The problem with genetic data isn’t just what people can do with it today, or what they can deduce about you today, it’s what they can do with it in the years ahead, as science evolves. France, to take one example, has sensibly outlawed such home DNA testing kits.

The company, 23andMe may finally die now. The CEO, the former wife of a Google founder, wants to buy it out of bankruptcy. But your most sensitive personal data, if you were stupid enough to spit into a test tube for them, will live on, no matter what happens to the company.

The AI training bots are reading my 100% human-generated blog…Great?!

2025-03-17T11:19:00.000+01:00

I have been posting to this blog to share my thoughts with a small community of privacy professionals.

So, I was a bit surprised to see Blogger give me statistics: my posts get around 10,000 views. I was surprised, because the privacy expert community is smaller than that.

But how many of those views were bots, in particular AI training bots? Blogger doesn’t give me those statistics.

We all know that AI models are trained on data. Big models, like large language models, are trained on vast amounts of data. In fact, they’re being trained on essentially all available data in the world. So, given their hunger for data, in particular for human-generated content, I’m not surprised they’ll visit my little blog too.

There’s a raging debate about whether AI training bots should be allowed to use other people’s data to train their models. There are many voices who claim that AI bots shouldn’t be allowed to train on other people’s data, if that data is either “personal data” or under copyright. I think they’re wrong.

I think the key distinction is public v private data. If I make my data public, as I do with this blog, then I should expect (and probably want) it to be read by anyone who wants to: humans or bots. After all, search engine crawlers have been crawling public data for decades, and almost no one seems to object. If AI training bots are reading my blog, say, to learn about human language, or about privacy, I’m delighted.

On the other hand, private data is private. If I use an email service, I expect that data to be private, as it’s filled with my highly personal and sensitive information. If I use a social networking service, and I set the content I upload to “private”, I expect the platform to respect that choice, including from their own or third-party bots. Failure to respect these privacy choices is a serious privacy breach, maybe even crime, unless the owner of the data has consented to allowing their data to be used for AI training. (It’s a different discussion if “consent” can be deduced from some updated clause in some terms of use.)

Thousands of training bots are looking for more data, especially human-generated data. If you make your data public, then realize the bots will come read it. You can’t really stop it. And I think that’s fine.

The real issue is what the AI models intend to do after training on your data. If they’re learning human language (large language models), it’s not going to have any impact on your real-world privacy. But if they’re reading your data to impersonate you, to copy your voice or image or your copyrighted content, then you have every reason to object and use the legal resources available. I think it’s fine when bots read public data for training. The real question, and vastly harder to evaluate, is what their trained models should be allowed to do with it afterwards.

Big and Rich v Small and Smart, who’ll win the AI race?

2025-03-10T13:03:00.000+01:00

Everyone is in the race for AI, in particular the race for AGI, artificial general intelligence. AGI was widely viewed as science fiction only a few years ago. Now the experts think someone will build AGI in the next year or two. Even if they’re wrong, AGI is coming soon. The consequences for homo sapiens are mind-boggling. https://www.livescience.com/technology/artificial-intelligence/agi-could-now-arrive-as-early-as-2026-but-not-all-scientists-agree

Who will win the race to build AGI? There are two camps. The first are the big, rich legacy tech leaders, often running large super-profitable monopolistic services that they are leveraging into the new world of AI. They have lots of advantages. Their businesses, or monopolies, generate vast amounts of money, user data, user engagement, installed bases. Legacy monopolies always use their core original monopoly to leverage into new businesses. For example, Google built a monopoly in Search, which was legal, but used that monopoly to leverage into many other businesses, according to the US Department of Justice. The US Department of Justice is bringing an antitrust case against Google and its abuse of its Search monopoly, but it has already decided to let Google continue to leverage that monopoly into the world of AI. https://nypost.com/2025/03/07/business/feds-drop-bid-to-make-google-sell-ai-investments-in-antitrust-case/ If that’s the best the Antitrust Division of the Department of Justice can do, well, I’d support the DOGE efforts to save taxpayer money and just eliminate them.

I assume they’re cracking open the Dom Perignon in Mountain View. The antitrust regulators in DC will let Google proceed with its old playbook. Maybe they’ll force Google to divest one of its portfolio of multiple monopolies, like Chrome. Big deal, that’s like pruning a branch off a tree. I imagine the company will howl with indignation, even at that low-impact antitrust remedy, but that’s like a dramatic husky howling at the vet.

On the other hand, maybe the old legacy tech world won’t win the AGI race. Maybe smarts, innovation, and agility will win. Maybe small companies and research labs will win. Maybe the future isn’t in the mega-model of AI, based on vast amounts of data, vast amounts of GPUs, vast amounts of money, electricity, and users. Maybe smaller models will win, figuring out how to do things on the cheap. The low-cost Chinese DeepSeek success, if it’s true, might be a window into that future.

In privacy terms, it’s not clear which model is better. The mega-model is based on vast amounts of data processing, limited to a few mega-companies. The size of the processing may be troubling, but it’s somewhat easier to hold big companies to account for responsible data processing. If the smaller models prevail, there will be a proliferation of AI processing across thousands, or maybe millions of actors. Good luck trying to ensure privacy rights in that scenario.

What do I think?: I think the smaller models will proliferate, eventually, after an initial lead by the big models. Fasten your seatbelt, we’re entering a zone of turbulence.

The leaning tower of privacy

2025-03-05T12:10:00.000+01:00

I’m now free to say what I think. For many years, I was an official spokesperson on privacy issues, amongst other things, for my prior employer. Naturally, I was committed to advocating for its views, as any good lawyer does on behalf of their client. I hope I lived up to my goal of only saying things that I believed were true, and not just parroting talking points.

Now that I’m free to say what I think in my personal voice, where should I do it? I’m on blogger.com, out of an old habit. Nothing much has changed on blogger, and it feels oceans apart from cool social media hotspots, but at least it’s familiar. What are my alternatives? As a privacy professional, I couldn’t possibly join FB as a “dumbf*ck”. I can’t join X, given how I feel about Elon. I can’t join TikTok, given how I feel about China. So, I guess I’m on blogger for now. Unless you have a better idea.

What I do care about is finding ways to share my experience, knowledge, and thoughts after 30 years’ of privacy practice, with students, privacy professionals, advocates, and regulators. For me now, it’s about sharing and helping a new generation in the field. I’ll be writing, speaking, teaching, advising, mentoring, as opportunities come around.

For example, why has the regulatory DPA world had such a limited impact on Big Tech that it is being asked to supervise and regulate? The DPA world has many strong tools, in particular, the tough law of the GDPR, but its toughness on paper hasn’t translated into the real world, as people expected when it was adopted. I could list some of the factors that have limited the DPAs’ ability to have a big impact. 1) The GDPR put first-line responsibility onto the shoulders of a one-stop-shop regulator, which turned out to be Ireland’s, for virtually all US and Chinese big tech companies. That’s a huge lift for the DPA of a small country, even if they have brilliant leadership and staff. 2) The DPAs have modest budgets, small teams, very few technical or legal experts, and they’re facing-off against mega companies with vast technical and legal resources. 3) The politics of being a DPA are complicated, since they are often accused of being retrograde, or anti-innovation, when they try to enforce the laws. 4) DPAs spend a lot of time on minor cases, often complaints by one individual, which might matter to that one individual, but have zero big impact. The Right to be Forgotten is a perfect example of individual-level cases that absorb DPA resources with virtually no impact beyond a particular case.

So, what’s my recommendation to DPAs to have more impact? Pick your cases wisely. Pick cases that affect millions of people, and don’t waste your resources on petty cases. Think about the tech and how it’s evolving, so that you don’t bring cases about 10-year-old tech that is already obsolete before any conclusion is reached. And spend time developing policy, at an international level, so that it’s clear what policy goals you’re pursuing. In particular, in the world of AI, push for international conversations and consensus on what good policy looks like in the world of AI, by engaging with stakeholders, and once that consensus is achieved (but not before), use your regulatory enforcement toolkit. I’ve become friends with many people in the DPA community. I trust them to want to do the right thing.

I could make the same recommendations to privacy activists: pick your cases wisely. The best, in my experience, in picking the right cases and pursuing them tenaciously, would be NOYB. He wouldn’t know it, but when I was on the opposite side, Max got me scared and sweating. I admire him for it. If you care about privacy, consider donating to NOYB.

Thou shalt not kill, unless thou art an autonomous AI killing machine

2025-03-03T11:53:00.000+01:00

I’m mostly bored with Google, but occasionally it publishes a blog that gets me riled up. When I still worked there, I was proud of my employer, and how it adopted a set of AI Principles, to govern its work in this exciting, promising, dangerous new space. One of those principles rejected work on AI that would “cause or are likely to cause overall harm.”

But guess what, it turns out that the militaries of the world want to harness AI for their purposes, which include “causing harm”, to put it mildly. So, Google, presumably smelling a big business opportunity of developing AI for the world’s “democratic” democracies, re-wrote its AI principles. Now the company can develop AI tools for the world’s democratic militaries (I’d love to read the list of the countries that Google considers “democracies”).

I re-watched Robocop recently: an AI-prototype robot cop is demoed in a boardroom, and a hapless person is asked to hold a gun pointing at it. Chaos ensues, as the prototype disregards human voice commands to stand down and shoots down the poor person holding the gun he’s trying to throw away. Oops. Indeed, the company whose AI gave us a recipe for glue on pizza, will now work on AI to kill people.

The collaboration between the militaries of the world and tech companies is wide and deep. Think space exploration or even the origins of the Internet. But here we’re talking about bringing very specific competencies of tech to build AI autonomous killing machines, building on these companies’ decades of leadership in surveillance, monitoring, profiling and targeting. It’s one thing, I think, to engage in surveillance, monitoring, profiling and targeting…to show ads, and quite another to select individuals or groups for elimination. The tech is not so different.

Any company can of course change its principles, or its ethics. Users can decide if they trust a company that changes its principles and ethics. Its workforce can decide if they want to work on these projects, even if they’re threatened with termination for refusing.

Privacy law, in Europe, for decades has included a principle that machines shouldn’t be allowed to make “automated decision making” on important questions without human involvement. Needless to say, an AI automated decision to kill someone seems to meet that test. The militaries of the world are largely exempt from privacy laws, but the private sector companies working for them are not. I would love to see people explore exactly what a company plans to do to build AI tools for the military, and ask the hard questions: how did you train it, what data did you use to train it, how reliable is it, what is your accountability for its accuracy or failure? You and I might think this is an intellectually interesting question, but we’re not teenagers on the frontlines of Gaza or Ukraine.

Surveillance just got a lot creepier

2025-02-24T12:06:00.005+01:00

I read a recent Google blog. Updating our platform policies to reflect innovations in the ads ecosystem. I have absolutely no idea what it was saying, which says something, considering I spent decades writing these blogs. If obfuscation was a literary genre, this was Shakespearian.

But other people figured out what it meant: Critics say new Google rules put profits over privacy Basically, Google was reversing its long-held pledge to fight the privacy-evil practice of fingerprinting. Most people are aware, or at least dimly aware, that they’re being watched, followed, profiled and targeted as they surf the web, especially by the ads ecosystem. After all, the more a company or an industry can monitor, profile and target individual users, the more money they can make from targeting them with individually-tailored ads. Virtually no one understands the scope and scale of these practices, including me.

As a privacy professional, I always defended a pragmatic approach. An ads-based ecosystem could co-exist with privacy laws in a cookie-based world, given that there were certain protections, transparency and user controls for cookies. And the regulators of the world largely approached online ads surveillance problems from a cookie-based perspective. But while they were barking up the cookie-tree, the industry was rolling out far more invasive and invisible individual-level tracking and profiling tools called fingerprinting. If you don’t know, fingerprinting is a technique to collect lots of individual little settings about users’ devices that can identify them uniquely, much like a fingerprint of a person’s thumb is composed of lots of little lines that individually mean nothing, but together identify a person. Fingerprinting is an evil privacy practice, with almost zero transparency or user controls. In my long privacy career, I always held the line against it. Increasingly, small players in the ads ecosystem resorted to fingerprinting, as an alternative to cookies. Google, after decades-long principled objection to it, has now raced to the bottom to join its competitors. Of course, it’s entirely different for the super-dominant player in the ads ecosystem to join its tiny competitors in bad privacy practices.

Perhaps privacy regulators will learn from their mis-guided focus on cookies to look at fingerprinting. But the USA is de-regulating fast, and Europe is out-gunned and out-manoeuvred by some players in the tech industry. Think about the resource imbalance, to take one public fact: Google’s CEO is paid, individually, about the same as the entire operating budgets of all 27 EU DPAs, combined. Now picture a small, valiant under-resourced DPA trying to take that on.

I think of privacy as a series of ethical choices to respect the individual, and laws that try to back that up, however imperfectly. But in the war of principles v profits, it’s hard for principles to win. Maybe shareholders will get richer, but humanity will be poorer. For me personally, it’s sad to see your life’s work deracinated.

I've left Google

2025-01-21T19:32:00.000+01:00

My nearly 2 decades at Google as its Global Privacy Counsel has ended. I’ve left Google as one of the last few remaining members of the original early Google team. Google asked me to update social media profiles accordingly, hence my coming back to this dormant blog to say I’ve left Google. Together with me, other senior members of the Google privacy team have left in recent months.

My career started as Google’s first full-time privacy professional, building a function, and later team, that didn’t exist before. My job was to try to make Google respect privacy for its billions of users. You can judge the results, but I am proud of the mission. Being a privacy leader is a tough job at a company like that.

The early years at Cool Google were fun, creative, innovative, comradely, and I loved them. But Google has changed and evolved into Corporate Google, and large committees can now carry forward the work I did, or reverse them; in either case, it’s no longer my business.

I left on good terms. No one is in jail now for privacy, including me, and I’m hardly being flippant, speaking as one of those rare privacy professionals who was arrested and sentenced to jail for their employer’s privacy practices. And I helped build the small company I joined into the largest private processor and monetizer of personal data on the planet. I can’t think of another privacy professional who helped build their data-processing company from the early days to 2 trillion + market cap. What a ride.

I will remain active in the field of privacy in many ways. AI will present existential challenges to the field of privacy, as to so many other domains, and I’m eager to find ways to help organizations develop AI responsibly. And there are innovators out there who remind me of the fun, creative, responsible environment of my early years at Google, as we wrestled with privacy issues and the then-new online world. Unless compelled by law to testify, I won’t reveal any non-public information about Google: I’ll respect my confidentiality constraints as a lawyer to my former client/employer.

I relish my newfound freedom to share my insights and experience with others in the field and in new ways. More on that soon. In the meantime, I wish luck to my former colleagues with MAGA: Make Alphabet Great Again.

Harvard Nostalgia

2014-05-14T08:56:00.000+01:00

The older I get, the more I think back to Harvard with nostalgia. I was very young at Harvard. I went to college straight out of the 10th grade, and I graduated at the age of 19. That officially makes me a high school drop-out.

I lived my teenage years in crimson, with the usual teenage crises, stumbling to figure out how to grow up faster than I should have. It was an exhilarating age of discovery and almost reckless ambition.

My years at Harvard had nothing to do with preparing for a job, not even at Harvard Law School. They were years of general education, following the classic liberal arts curricula. I went to bed almost every night with Shakespeare. Harvard is where we were learning hard lessons, groping to try to be masters of our own fate.

I have a few regrets. I wish I had stayed at Harvard longer. Why did I feel it time to graduate at 19? Why didn't I just pick another subject or another degree or start a business in my dorm? Why didn't I just take more time then, when the world was on my platter, rather than rush into the long muddle of middle age?

As someone who was in college in the early 80's, there are hardly any records or momentoes left to reflect on that time. Hardly any photos of friends or places. No tweets or blogs to re-discover and remember. My personal historical archive is bare, compared to kids' today. Sure, there are no embarrassing photos on the web and no one had even heard of the concept of cyberbullying, but then again, all the rest of life memories have largely evaporated, like a decaying Widener of the mind.

The most important parts of my moral compass were set at Harvard, in particular, the sense of privilege and responsibility for belonging to an obvious elite, where it was just natural for classmates to become Nobel Prize winners or tech billionaires or poets, or mediocrities bedevilled by a nagging sense of unfulfilled promise. Even a classmate who becomes President is judged as a disappointment, based on a sense of promise unfulfilled.

There's time left, I tell myself, time to shake it up, before the sum-up, before those pithy obituaries in the Harvard Magazine, like the usual ones: so-and-so died suddenly while fly-fishing in Patagonia after a career in law, survived by his spouse (Harvard Class of XX), and leaving his modest estate to fund a scholarship for swimmers at his Alma Mater.

I want to go back to Harvard, not to some 30th reunion, but metaphorically, to that time of endless opportunities, where Gates and Zuck were gestating, where Obama polished his law-professor-with-politician's-smile, where Yo Yo Ma dashed down the hall with his cello case, and where the best parts were private. Thirty years later, I walked down the dilapidated halls to give a lecture at an "elite" German university, with its egalitarian-ethos and 100,000 or more students, and I thought back to my time at Harvard, and whispered to myself once again, we few, we happy few.

When everyone else seems to be trying to figure out how to delete and edit their life histories, or at least the public fiction of their life histories, I'm fumbling to hang on to mine.

From pool to pool

2014-04-06T14:07:00.000+01:00

I love my pools.

In Paris, my pool is in a fancy private club, in the Bois de Boulogne, on the edge of Paris. It's a gorgeous 50-meter pool, open year-round outdoors, framed by views of the Eiffel Tower. My pool nurtured many Olympians, starting a century ago. Nowadays, it's mostly rich old people, who seem vaguely annoyed with me, a hard-swimming American. At my French club, we don't admit many new members, we keep the gates high for privacy, and we don't really want anything to change.

In Ft Lauderdale, where I grew up as a kid and where my parents now live, I swim in a historic 50-meter public pool, located feet from the beach. It was one of the first 50-meters in the US, dating back to the 30's. Nowadays, it mostly hosts visiting swim teams from around the world. It's fun for me to swim with college teams. They're young and strong, and they swim with modern techniques. I enjoy the purity of it. You can't fake swimming. No billionaire can buy a butterfly stroke. And here, for two bucks, you can swim under the Florida sun, breathing the salty air, and swim with the guys from Calgary on Monday, and the guys from Bologna on Tuesday, on this spot where people have been re-inventing the sport for 80 years, and the handsome smiling Italian in the next lane tells me he loves America.

And then there's Blodgett, the pool at Harvard, the most exclusive of them all. I was so intimidated and excited. How exhilarating, at that age, for the first time in my life, to be meeting Harvard guys with high IQs and low times. And I still wake up at the side of one of them every morning.

From the outside, swimmers are easy to spot. We reek of chlorine, we have bad hair, we get up at 5 to hit the pool, we sacrifice evening social events, we slouch around like exhausted zombies.

But when I slip into the water in the morning, I feel like I'm finally coming alive again. The rhythms play and change constantly, the endless counting, the new goal every 30 seconds, visualizing each rotation, each flip, each catch as the first chance to get it just perfect, after a million tries.

It's the same when I swim with a team and when I swim alone. The internal pressure is the same, the mental games, the exertion, the exhaustion, the elation. I can still remember swimming with my first team, as a little kid, and trying, over and over again, to learn the flip turn, like some deranged hamster.

To call it discipline doesn't capture it. I wake up with shoulders so sore that the thought of swimming makes me want to cry. But an hour later, that's exactly what I'm doing, pushing this fragile and tired swimmer's body through the water.

What a gilded life, to spend my teenage years wandering around the Harvard campus, carrying a swim bag from lecture hall to pool. At that age, my father was wandering around Berlin, a Jewish kid given the job of picking up unexploded bomb shells, a child forced to carry death in a wheelbarrow.