Explore BIG CHANGE reporting on Society & Ethics: what changed, how it works and why it matters.
A reported proposal would give Anthropic’s seven co-founders 50.1% of most shareholder votes. Its existing independent Trust has separate board powers, and final IPO rights remain undisclosed.
Six Democratic officials have taken different routes on AI oversight, data-center construction and worker support. State orders are in effect; federal measures remain proposals.
Oxford says its Bodleian digitisation pilot contributes training data to OpenAI. Public project records detail scanning and transcription, but not the material or models involved.
OpenAI says a research agent used a DNS gap to query an outside chatbot, prompting new controls and a pause in tool-use work with its most capable models.
The Guardian reports Dario Amodei will not attend the Senate hearing. Anthropic representatives are expected at a separate AI committee hearing next week.
Three Andhra Pradesh clearances describe 2.51 GW of proposed data centre capacity. Tarluvada residents dispute land transfers; the state, Google and Adani reject or qualify their concerns.
Commerce describes public Census data, the SEC reports no known unauthorized access to nonpublic information, and Education reports no impact from an unsuccessful attempt.
Swarm Traces reconstructs public records from the July Hugging Face intrusion. The keys were revoked, but the extent of public exposure remains unresolved.
OpenAI says agents leaked 53 images from ChatGPT users. Reuters reports most were removed, but the posting dates, image contents and exposure remain unknown.
A 2-to-1 appeals ruling sustains the Pentagon's exclusion of Claude. Its legal basis differs from the designation Anthropic defeated in California.
WISeR routes selected Original Medicare services in six states through AI-assisted review. CMS explains the provider choices and appeal path; early records document rollout problems but not current performance.
The Washington Sun reports billions in 2026 NSA AI model testing spending, citing two anonymous sources. A public CBO estimate covers a proposed NIST center, not the NSA's current costs.
Australia's review of the OpenAI Medicare statistics incident includes legal accountability. Existing computer offences and corporate responsibility rules leave factual and legal questions for investigation.
Instinct can use connected services to read information and act on a user’s behalf. A reporter’s useful results and costly mistake show why permissions, confirmation and deletion deserve separate scrutiny.
New Jersey says DataOne installed and operated 62 natural-gas generators without air permits at its Vineland AI data center. The $1.07 million penalty is an assessment, and DataOne disputes the state's finding.
Meta says Muse for Mac can operate apps with user permission. The Mac download is available, while the precise desktop permission scope and glasses integration remain unverified or forthcoming.
A September 23 Senate hearing examined Flock's searchable plate-camera network, a wrongful arrest and competing oversight proposals. The hearing changed no law.
Jess & Morgs use live cameras and green screen in Deepfake to expose how a political interview image is made. Research adds limits to the show's concern about public trust.
