Former OpenAI safety researchers ask oversight bodies to protect outside collaboration
Three dismissed OpenAI researchers urge oversight bodies to protect outside safety work and clarify employee rules. OpenAI disputes their account of the firings.
Explore by topic
Society & Ethics
Explore BIG CHANGE reporting on Society & Ethics: what changed, how it works and why it matters.
Latest reporting
115 stories
Anthropic’s Usage Policy update takes effect November 12
Anthropic says its November 12 Usage Policy update adds a narrow ban on sustained, needless abuse toward its models and safety requirements for hardware connected to potentially injurious autonomous actions. Most other revisions clarify or reorganize existing rules.
Germany’s cabinet backs bill covering sexualized deepfakes and wider digital abuse
Germany’s cabinet approved a digital violence bill covering sexualized synthetic imagery, voyeurism and other abuse. Parliament must still consider the proposal.
Michael Smith sentenced to 18 months for bot streams of AI-generated music
Michael Smith received an 18-month sentence after admitting to a wire-fraud conspiracy built on fake music streams. AI-generated songs supplied catalogue volume, while automated accounts produced the plays that generated royalties.
How to check an image, video or audio file with Google SynthID Detector
Google has opened SynthID Detector globally in English for image, video and audio checks. This documentation-based guide explains the stated workflow and what a detected or missing watermark can establish.
Common Sense Media tests the limits of ChatGPT’s teen safety alerts
Newly linked teen test accounts received no parental alerts during scripted crisis conversations, the Youth AI Safety Institute reports. OpenAI disputes the test timing.
Hark Pro’s Australian App Store listing shows an assistant built around connected accounts
Hark Pro is listed in Australia’s App Store. Its website, store listing, privacy policy and terms describe different parts of access, pricing and data handling before a user connects accounts.
Finnish regulator asks Tuike to suspend specified work at Google's Muhos and Kajaani data center sites
Finland's environmental regulator asks Tuike Finland Oy to suspend environmentally significant preparations at Google's Muhos and Kajaani data center sites while their impact assessments remain unfinished.
McDonald’s faces lawsuit over alleged AI-assisted price coordination
A proposed antitrust suit alleges McDonald’s pricing recommendations helped coordinate restaurant prices. The docket verifies the filing, while the company says franchisees set prices independently.
Pinterest Beauty Guides turn hair and nail Pins into saved references
Pinterest says Beauty Guides turn hair and nail Pins into planning references that users can save. The launch is available to U.S. iOS users; the company has not published an accuracy assessment or a price.
ChatGPT cartoons carried real New Yorker cartoonists’ signatures
Nieman Lab found ChatGPT-generated cartoons bearing real New Yorker cartoonists’ signatures. Its tests show false attribution, while the training source and rights remain unresolved.
Utah agreement permits Nolla AI acne prescriptions; pilot start unconfirmed
Utah's agreement permits Nolla to initiate a narrow set of topical acne prescriptions for adults, but the state says the demonstration has not started while Nolla says it is available.
IWF counted 6,310 AI-generated child sexual abuse images in six months
IWF assessed 6,310 AI-generated child sexual abuse images in the first half of 2026. Its conflicting 2025 totals leave the exact growth rate unresolved.
Preliminary report links Amap agent queries to Tencent Cloud infrastructure
Researchers report parallel AI agent queries for Amap entrance-navigation data and traces tied to Tencent Cloud infrastructure. The operator and purpose remain unknown.
Norway plans a temporary AI glasses ban in selected places
Norway's government plans to propose a temporary AI glasses restriction in selected places. The devices, locations and exceptions remain unsettled.
AI memory demand is squeezing budget smartphone access
Retail price rises and falling shipments in cheap phone bands show how memory pressure affects handset affordability, with unequal consequences across regions.
Arizona court vacates sentence over AI victim video
An Arizona appeals court affirmed Gabriel Horcasitas's manslaughter conviction and ordered resentencing over an AI recreation of the deceased victim.
How to use Gemini Live Guided Vision to read a small label
Google's Guided Vision can describe what an Android camera sees. Follow the documented steps to enable it and ask about a small label, with its safety limits in view.