Vincent Toubiana

At 2pm Paris time today I'll be testifying in the French Parliament about digital sovereignty.

Watch it live at videos.assemblee-nationale.fr/direct.php!

Metacritic Removes Resident Evil 9 Review From Fake AI Writer Videogamer's human staff was wiped out and replaced with AI slop

Quelle angoisse : le site britannique Videogamer a licencié tous ses rédacteurs la semaine dernière, et depuis le site ne publie plus que du contenu IA (avec de faux profils de faux rédacteurs), dont des tests de jeu ???

kotaku.com/resident-evi...

📢 The 5th Privacy Research Day will take place on June 24
📅 The call is open until March 22

Topics:
🧒 Vulnerable Users
🗺️ Internationalisation of privacy
🤖 AI
🕶️ Wearables/IoT
🧑‍🔬 Science & regulation
💳 Digital economy
🛡️ PETs
🔓 Usable Security & Privacy

More information: www.cnil.fr/en/call-pape...

Let's try this again 👋

Hi Bluesky - I'm Massachusetts' Attorney General. You might know me from suing President Trump nearly 50 times, beating Uber and Lyft in court, or being the first woman of color elected to statewide office in MA.

I officially left X today - help me find my MA people?

Finally a home for Grok-generated CSAM the EU doesn't want you to see

Google "provided ICE with the usernames, physical addresses, and an itemized list of services associated with the Google account of Amandla Thomas-Johnson, a British student and journalist who briefly attended a pro-Palestinian protest in 2024 while attending Cornell University in New York"

One Tap To Hijack Them All - A Security Analysis of the Google Fast Pair Protocol YouTube video by COSIC - Computer Security and Industrial Cryptography

🎧COSIC reveals hundreds of millions of earbuds & speakers need a patch to prevent wireless hacking & tracking. Flaws in 17 models using #Google’s #FastPair let attackers hijack devices, eavesdrop, and track users.
Watch our demo: www.youtube.com/watch?v=-j45...
#cosic #kuleuven

This thread is a very nice teaser :) Hard to find a good and effective solution to this problem...

My point here is that many of the data brokers trafficking in consumer data are not registered in California, many are fly-by-night operations, and frankly, aren’t paragons of integrity.

Asking—only the ones known about—to delete data post hoc is closing the barn door after the horse has left.

9/

We were able to reidentify 97% of the individuals in the dataset (showing it’s not anonymous).

None of that required IRB approval (i.e., humans weren’t the subject of the experiment).

Emailing the reidentified humans to ask about their recollections of consenting to these sales, however, did.

3/

Delete request and opt-out platform (DROP) Protect your personal information. Data brokers collect, share, and sell your personal information. You can stop that from happening.

Those in California can now go here to request that data brokers delete their data: privacy.ca.gov/drop/

This is a good thing, however, it’s far from a complete solution. 🧵

1/

👻 Avez-vous téléchargé votre FantomApp ?
La 1ère application de la #CNIL pour aider les 10-15 ans sur les réseaux sociaux : des outils pour protéger ses comptes et nettoyer sa présence en ligne, des conseils en cas de problème, une application qui ne collecte aucune donnée.
fantomapp.fr/home

Imagine the possibilities. Such an opportunity for everyone from meddling foreign intelligence to your bog standard chaos agent, with zero friction.

Looks like Waymo cars are trying to drive more like humans drivers. Not sure it's a good thing

techmeme.com

N.Y. Law Could Set Stage for A.I. Regulation’s Next ‘Big Battleground’

New N.Y. law requires retailers that use personalized pricing to post the following disclosure: “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA.”

www.nytimes.com/2025/11/29/n...

How we’re responding to The New York Times’ data demands in order to protect user privacy OpenAI is fighting a court order at the demands of The New York Times and plaintiffs, which involves retention of consumer ChatGPT and API user data indefinitely. Learn how we’re working to uphold use...

"After months of litigation, we are no longer under a legal order to retain consumer ChatGPT and APl content indefinitely. Our obligations under the earlier order ended on September 26, 2025."

openai.com/index/respon...

Meta is earning a fortune on a deluge of fraudulent ads, documents show Meta projected 10% of its 2024 revenue would come from ads for scams and banned goods, and it internally estimates that its platforms show users 15 billion scam ads a day, company documents show.

"Every six months, Meta earns $3.5 billion from just the portion of scam ads that 'present higher legal risk,' the [Meta] document says, such as those falsely claiming to represent a consumer brand...That figure almost certainly exceeds 'the cost of any regulatory settlement involving scam ads.'”

Picture from Google blog announcement about Gemini integration into Google maps. I highlighted the sentence "We’ve all heard “turn right in 500 feet” while driving and wondered exactly how far that is. "

No we have not. Most of us use the great, precise, yet very clear, metric system and know exactly when we must turn right.

Only the US, Liberia and Myanmar use the "Imperial system" and rely on traffic lights and restaurants to know where they should turn😏

blog.google/products/map...

I cannot emphasize enough how much the folks at Census are *obsessed* with doing the right thing for both representative data collection and privacy. (I say this as someone who nearly didn’t get to tenure because of all the frictions involved in using restricted-access Census data.)

🗨️ "Austrian privacy group noyb said on Tuesday it has filed a criminal complaint in Austria, accusing U.S.-based Clearview AI of illegally collecting photos and videos of European Union residents to build its facial-recognition database."

More Info:
www.reuters.com/sustainabili...

Wouldn't it be possible to detect META glasses based on their MAC addresses? I assume they are using the same vendor prefix

Another solution would be to broadcast a BT signal when they record so that people could at least be informed (using Apple & Google anti-stalking alerts if that's possible?)

This idea from @esquiring.bsky.social is totally doable but has minimal chance of being embraced by Meta.

Sur ce sujet, lisez le dernier Cahier Innovation et Prospective de la CNIL, sur les données post mortem et notamment les "deadbots" :

Article : linc.cnil.fr/cahier-ip10-...
Pdf : linc.cnil.fr/sites/defaul...

To call this feature an "URL shortener" when in many instances it makes URLs longer is, at least, a bit deceptive. Not sure this is completely legal

They could at least check the new URL is not longer

"The watchdog said the option to keep the cheaper plan was only revealed after consumers began the cancellation process, a design it argued breached Australian consumer law by failing to disclose material information and creating a false impression of available choices."

Thanks!

I thought there could be some rationals beyond "collecting more data". There are several things that I find very inconvenient and/or misleading.
First, in many instances the *URLs are not shorter*, so to call that a shortener is misleading.
Second, all shortened URLs look the same... [1/3]

Brought to you by the company who discontinued its "goo.gl" URL shortener in 2019 and killed it in 2025.

Fourth, it is not clear to me why this is a "Google Account" setting and not Chrome or Android setting.

Finally, when a default setting is changed, users should not have to go through "Other settings" to restore the original configuration. At least not immediately [3/3]

It is problematic both for the person sharing the URL (who may not notice he/she share the wrong link) and for the person clicking on a link who does not know where the link is redirected.

Third, adding an unwanted intermediary service adds delays and a point of failure.[2/3]