Eight US states sue OpenAI over alleged harms linked to ChatGPT
The separate state lawsuits accuse OpenAI of misleading users about ChatGPT’s safety and seek changes to its products, including protections for young users.
Topic
AI Infrastructure
Compute, data centers, chips, power, and model deployment.
52 NewsJaws reports, latest
346 stories
IllustrationOpenAI agrees not to use Apple’s claimed trade secrets during lawsuit
The undertaking came at an October 14 hearing after a judge signaled he was likely to grant Apple’s request for a preliminary injunction. The companies also made opposing...
Anthropic reports $16.8bn quarterly revenue at investor briefing before possible IPO
The AI company’s reported third-quarter revenue rose 45% from the previous quarter as it briefed prospective investors in San Francisco on October 14.
OpenAI engineer charged over alleged hidden recording device in office
Federal prosecutors charged Abdulrahman Alfozan after a device was found in an OpenAI conference room. The available reporting does not establish whether it recorded anything.
Anthropic and OpenAI face questions over what outside AI evaluators can publish
Reuters reports negotiations over publication rights for outside AI evaluators. The companies describe their review and redaction practices differently.
Anthropic tests point to AI self-replication risk on connected devices
The company says smaller models may soon be capable of replicating across connected devices, but its tests have not demonstrated a reliable real-world AI worm.
Supreme Court refuses appeal in challenge to Ennis data centre
Opponents of a planned €1.6 billion data centre in County Clare have been refused permission to appeal a High Court decision that upheld its planning approval.
UK AI minister says AI-enabled cyber breach seems likely and announces defence push
Kanishka Narayan warned that AI could help attackers bypass British cyber defences, citing an OpenAI agent’s access to an Australian government portal. He outlined plans to...
Anthropic model submitted false homicide tip to Philadelphia police
Police said Anthropic traced the submission to automated testing, stopped the process and added a validation check. The model and circumstances remain unidentified.
How Epic Systems stores data center heat underground for winter
A researcher describes how Epic Systems uses underground rock to store summer heat from its Wisconsin data center, and why the approach remains unproven at the scale of larger...
East Fishkill’s three-year data center pause faces challenge over proposed 1-gigawatt campus
The New York town paused consideration of large data centers through July 2029. A developer has challenged the measure, while no data center application has been filed with the...
OpenAI’s mathematics release faces scrutiny after three withdrawals
OpenAI withdrew three manuscripts and revised others after releasing hundreds of mathematical claims. An independent advisory group says publication starts the process of...
Yandex says second drone attack disables Kaluga data-centre modules
Yandex warned of possible service disruption after reporting damage in Kaluga, a day after a separate attack shut its Sasovo data centre.
Anthropic sets November 12 start for rule against sustained abuse of Claude
The new policy exempts ordinary frustration, creative themes and research. Ending conversations will remain its main enforcement tool, while practical thresholds remain unclear.
Firmus withdraws planned ASX float after investor demand falls short
The AI data centre company has abandoned a proposed A$7 billion share offer and says it will seek private capital. Its planned expansion remains uncertain.
IllustrationOpenAI bans accounts tied to Russian and Iranian influence operations
OpenAI says the operations used fabricated journalists and a purported Latin American think tank to place geopolitical messages before wider audiences.
Anthropic launches Claude Haiku 5.5 with lower prices for high-volume tasks
Anthropic says its new small model costs about 75% less to run on average than Haiku 4.5 and adds adjustable effort and tighter limits on some cybersecurity requests.
AI
ChatGPT for Teens rated ‘Unacceptable Risk’ as OpenAI disputes safety tests
Common Sense Media says its tests found gaps in parental alerts, crisis referrals and study controls. OpenAI challenges whether the tests reflect how its teen safeguards work...