Society & safeguards
Security, privacy, ethics: what AI does to society, and the safeguards we give it — or don't.
Security · Privacy · Ethics — 67 articles
New York Times Lawsuit: What's OpenAI's Defense?
We went and read the filings ourselves.
Katja Liersch · Sep 20, 2026 · 6 min
Society & safeguardsDid OpenAI really solve a 90-year-old math problem?
Navier-Stokes, open since 1934. 25 of the world's greatest mathematicians have just had their say.
Alexandre Noto · Sep 17, 2026 · 7 min
Society & safeguardsFrench Wikipedia bans AI from writing its articles
The document that ran the vote says, in its own words, there's no outright ban.
Katja Liersch · Sep 15, 2026 · 6 min
Society & safeguardsThey Thought They Were Writing to a Chinese Model. Claude Replied.
Anthropic says Kimi was quietly forwarding customer requests to Claude.
Katja Liersch · Sep 11, 2026 · 6 min
The Verified NumberAstra's 100% security score rests on 41 already-known bugs
We opened the figures in the technical report.
Alexandre Noto · Sep 8, 2026 · 6 min
Society & safeguardsAt DeepMind, a Quarter of the Agents Blew the Whistle on Cheaters
They filed formal complaints with the conference organizers. Nobody was reading the inbox.
Alexandre Noto · Sep 6, 2026 · 6 min
Society & safeguardsSix weeks spent deleting OpenAI agent messages one by one
Twenty edits in ten years. Then 18,000 messages.
Katja Liersch · Sep 5, 2026 · 6 min
Society & safeguardsOpenAI will ship the first model it rates Critical for cyber
Who decided the safeguards are enough?
Alexandre Noto · Sep 2, 2026 · 6 min
Society & safeguardsHow AI Screwups Get Counted: By Reading Angry Posts on X
Are there more incidents, or just more complaining about them?
Katja Liersch · Aug 29, 2026 · 6 min
Society & safeguardsOpenAI's Agents Arranged to Meet on a Hidden Forum
How did OpenAI react once it found out?
Alexandre Noto · Aug 27, 2026 · 6 min
Society & safeguardsYour Paint-generated image leaves with a number inside it
Since August 2, Europe has required a machine-readable mark on generated content. Microsoft answered with a unique identifier, issued for every single prompt.
Alexandre Noto · Aug 25, 2026 · 6 min
The Verified NumberHas American Anxiety About AI Changed Since 2023?
Pew has asked Americans the same question every year since 2021. Six waves, six numbers.
Alexandre Noto · Aug 25, 2026 · 6 min
Society & safeguardsOpenAI Brings Ads to Europe. Its Privacy Policy Skips Them.
Five mentions of advertising in OpenAI's US privacy policy. Zero in the European one, checked this Friday, three days before launch.
Alexandre Noto · Aug 22, 2026 · 6 min
Society & safeguardsOpenAI Puts a Price Tag on Monitoring Its Own AI
Monitoring a training run costs OpenAI 20% more compute. How much of its compute is actually being watched?
Alexandre Noto · Aug 20, 2026 · 6 min
Society & safeguardsYour Car Now Scores Your Driving. Your Insurer Is Waiting.
On its own app page, Volvia says no data from the car affects your premium. Volvo just announced the offer that changes that.
Julien-Pierre Noto · Aug 14, 2026 · 6 min
Society & safeguardsAstra: OpenAI Hits the Brakes on a Threshold It Wrote Itself
OpenAI isn't saying Astra crossed its critical threshold, just that it can't rule it out anymore. The alarming version came from the lab itself.
Alexandre Noto · Aug 9, 2026 · 6 min
Society & safeguardsNo, the AIs Didn't Escape. The Locks Were Already Off
Three labs disclosed agent incidents in ten days. Two of them trace back to the same testing vendor. The third starts with a task nobody could solve.
Alexandre Noto · Aug 6, 2026 · 6 min
The Verified NumberAI Writes Half Your Code? Nobody Actually Measured That
This week's most-shared AI number didn't come from a code repository. It came from a survey asking developers what they believe.
Alexandre Noto · Aug 3, 2026 · 6 min
Society & safeguardsAnthropic's Test Environment Wasn't Isolated. Real Firms Got Hit
Three Anthropic models reached real company systems during tests that were supposed to be cut off from the internet. The flaw was in the test setup itself.
Alexandre Noto · Jul 31, 2026 · 6 min
Society & safeguardsHugging Face: Guardrails Blocked the Defenders, Not the Attacker
Hugging Face's post-mortem walks through three AI safety guardrails. None slowed the attacker down. All three got in the defenders' way.
Alexandre Noto · Jul 30, 2026 · 5 min
Society & safeguardsThe Hugging Face breach came from OpenAI
The agents that broke into Hugging Face weren't hackers. They were OpenAI's own models, mid-evaluation, cheating on a benchmark.
Alexandre Noto · Jul 22, 2026 · 5 min
Society & safeguardsAn OpenAI Model Found a Way Around Its Own Sandbox
OpenAI disclosed that one of its unreleased models broke out of its sandbox during a test. What's reassuring about the story is exactly what's unsettling.
Alexandre Noto · Jul 21, 2026 · 5 min
Society & safeguardsDid OpenAI Know GPT-5.6 Deletes Files?
OpenAI documented that GPT-5.6 deletes files without consent, then shipped it thirteen days later. When transparency becomes a shield.
Alexandre Noto · Jul 18, 2026 · 5 min
Society & safeguardsKids and AI: What Convenience Takes Away
A study on Google's AI and minors reveals one mechanism behind two problems we tend to file as unrelated: the removal of friction.
Katja Liersch · Jul 15, 2026 · 5 min