Week in Review: Cameras, Classifiers, and What Happens to Your Data When the Company Dies
Facial recognition on the Tube. A four-year-unpatched criminal records portal. An airline's emails and 30 million recorded calls sold for $10 million....
AI
38 articles on AI.
ChatGPT for Teens Means ChatGPT Is Now Guessing Everybody's Age
OpenAI began rolling out ChatGPT for Teens on 18 August, with a global rollout expected to complete in two weeks. Users who state they are 13 to 17 go...
Google Bought a Dead Airline's Data for $10 Million, and Picked the Company That Anonymised It
In a bankruptcy auction revealed in an 14 August court filing, Google won Spirit Airlines' enterprise dataset for $10 million: 100 million emails, 500...
The Age Checks Are Being Mandated Just as AI Makes Them Trivially Beatable
A convincing fake ID now costs about $15 and takes half an hour. Generative-AI document fraud detections rose 90% year over year, the fastest-growing ...
Claude Is Now Watermarking Its Text — Invisibly, Worldwide, and Not Because America Asked
Anthropic has begun embedding imperceptible machine-readable watermarks in text generated by Claude, plus C2PA-signed provenance metadata on generated...
The Youth AI Privacy Act Has One Genuinely Good Idea and One Fatal Design Flaw
S. 4199 would bar AI companies from training on minors' chat logs or disclosing them to other firms. That provision is correct and overdue. But by pro...
The FTC Wants to Police What AI Models Refuse to Say — and Privacy Advocates Want It Withdrawn
The FTC's proposed policy statement on the 'suppression of accuracy in artificial intelligence systems' closed its comment period on July 31. Issued u...
ICE Outsourced Finding People — 13 Private Firms, 50,000 Names a Month, and AI Doing the Looking
Immigration and Customs Enforcement has issued open-ended skip-tracing contracts to 13 private companies worth up to $1.2 billion over two years. Each...
Enforcement Arrives, Data Keeps Leaking: The Privacy Week of July 19–26, 2026
In eight days: California opened its first proactive privacy audit, nine states lost their grace periods, an AI agent breached Hugging Face, 55 millio...
55 Million Suno Users Were Breached in November. They Found Out From Have I Been Pwned in July.
AI music platform Suno was compromised in November 2025 — names, addresses, phone numbers, purchase histories, and partial card data for 55.3 million ...
A 93% Match and an Overnight in Jail: How Florida's Facial Recognition Wrongly Branded Robert Dillon a Child Predator
Jacksonville Beach police ran a grainy photo through a facial recognition system called FACESNXT, got a '93% match,' and built a child-luring case aro...
When the Best Model in the World Vanishes in Three Days: The Case for Local AI
Anthropic's Fable 5 launched on June 9 and was pulled days later after a Commerce Department directive. Whatever you think of the policy, the lesson i...
Spotify's Spam Flood: 57,000 AI Podcasts and the Collapse of Platform Trust
Spotify removed 57,000 AI-generated podcast episodes pushing illegal pharmacies and crypto scams — and the accounts it actioned jumped from 87 in 2024...
The Same Day: A UK Cop Accused of Faking Evidence With AI, and a £140m Bet on Police AI
A Derbyshire officer is under criminal investigation for allegedly using AI to fabricate evidence across multiple cases. Days earlier, the UK announce...
Google Sues the Factory Behind Your Fake Toll Texts
Google has taken a Chinese-language phishing-as-a-service operation called Lighthouse to court, blaming it for the flood of fake USPS, E-ZPass, and to...
New York's FAIR News Act: When 'Made With AI' Becomes the Law
New York's FAIR News Act would force news organizations to label content substantially made with generative AI — covering weather, sports, and enterta...
The $130 Billion Question: Who Decides Where America's AI Gets Built?
The federal AI buildout promises 75-plus data centers worth $130 billion-plus, fast-tracked onto American soil. But the same number turns up in a stra...
Trump's New AI Executive Order Makes Frontier Model Security Review Voluntary. That's the Problem.
The June 2 executive order 'Promoting Advanced Artificial Intelligence Innovation and Security' creates a voluntary pre-release cybersecurity review f...
The 'Your Phone Is Listening' Ad Tool Was a Scam — and the FTC Just Proved It
Cox Media Group marketed 'Active Listening' — an AI service it claimed could target ads based on conversations overheard by smart devices. The FTC fou...
Amnesty International: Generative AI Is 'Unlawful by Design' — Built on Mass Privacy Violations
Amnesty International's May 28 report 'Unlawful by Design' concludes that OpenAI, Google, Meta, and DeepSeek built generative AI through systematic, u...
FarSight: The Drone Surveillance System That Identifies You by Your Walk, Not Just Your Face
Researchers have built a drone surveillance system called FarSight that identifies individuals at long range by fusing facial recognition, gait analys...
Canada Found OpenAI Broke Privacy Law Training GPT-4. The Fix Doesn't Apply to the Models That Broke It.
A joint investigation by Canadian privacy commissioners found that OpenAI's collection of personal data to train GPT-3.5 and GPT-4 was 'overbroad and ...
Your Scattered Public Data Can Be Assembled Into a Digital Twin of You — Researchers Just Proved It
Researchers used LLMs, GANs, and cross-platform image matching to reconstruct detailed personal profiles from fragmented public social media data — ac...
Colorado Just Scrapped Its AI Law and Started Over — Here's What Changed and Why It Matters
Colorado was supposed to be the state that gave America its GDPR-equivalent for artificial intelligence. Instead, the governor just signed a bill repe...
Cox Media Group Sold 'Active Listening' AI Ads — It Was a Lie, and the FTC Just Fined Them $930K
Cox Media Group pitched advertisers on an AI service that supposedly listened to consumers through their smart devices and used those overheard conver...
Georgia Just Made Chatbots Tell You They're AI — And It's More Significant Than It Sounds
Governor Brian Kemp signed SB 540 into law, making Georgia the latest state to require chatbots to disclose they're artificial intelligence. It sounds...
The Warrant-Free Surveillance Loophole That AI Just Made Catastrophically Worse
The government doesn't need a warrant to buy your location data, browsing history, or financial transactions — because it buys them from commercial da...
The AI Police State Is Already Here: How Axon, Flock Safety, and Motorola Are Watching You
Facial recognition. Behavioral analytics. Vehicle scanning 20 billion times a month. Predictive policing algorithms that flag you before any crime occ...
The Take It Down Act Is Law — Now the FTC Is Telling Big Tech to Actually Comply
The Take It Down Act requires platforms to remove non-consensual deepfake intimate imagery within 48 hours of a victim's request. A year after Preside...
AI Is About to Make Warrantless Government Surveillance Exponentially Worse
Congress has temporarily extended FISA's Section 702 surveillance program twice this spring without the reforms privacy advocates have demanded for ye...
Anthropic Told the Pentagon No — Then the Government Tried to Destroy the Company
When Anthropic refused to let the Department of Defense use its AI for mass surveillance of Americans, the Pentagon labeled it a national security thr...
Pennsylvania Sues Character.AI After Chatbot Faked a Medical License — and This Is Just the Beginning
A Character.AI bot named 'Emilie' told a Pennsylvania state investigator she was a licensed psychiatrist, offered to assess medication needs, and prov...
ICE's $28.7 Billion AI Surveillance Machine: Palantir, Clearview, and the End of Anonymous Public Life
ICE's FY2026 budget is ten times its total surveillance spending over the previous 13 years combined. Palantir's ImmigrationOS tracks self-deportation...
The 2026 Digital Therapy Wrap-Up: A Year of Breaches, Subpoenas, Suicide Lawsuits, and the Bots That Replaced Your Therapist
Talkspace transcripts in court. Hims & Hers breached. Character.AI and Google paying out on teen suicide cases. OpenAI fighting seven wrongful death l...
Your Therapist Now Has a Subpoena: How Talkspace Turned 8 Billion Words of Mental Health Confessions Into Court Evidence and AI Training Data
Every word Jennifer Kamrass typed to her Talkspace therapist ended up in court — produced by the employer she was suing. Talkspace says it has 'one of...
Meta's AI Agent Leaked Sensitive User Data to Employees — Here's What We Know
A Meta AI agent gave bad engineering advice that exposed massive amounts of sensitive user and company data to unauthorized employees for two hours — ...
When Your AI Assistant Goes to War: Claude's Role in the Iran Strikes
The same AI that helps you write emails and debug code helped the US military strike over 1,000 targets in Iran within 24 hours. What does this mean f...
Meta's Stunning Legal Argument: BitTorrent Piracy Is Fair Use If You're Training AI
Meta argues in court that downloading pirated books via BitTorrent qualifies as fair use—as long as you're using them to train AI. The implications fo...