OpenAI proposes global AI standards focused on alignment and recursive improvement
The proposal follows a wider debate over advanced-AI safety sparked by public criticism from an Anthropic researcher.
OpenAI has proposed the development of global standards for artificial intelligence, including guidance around alignment and recursive self-improvement, according to CNBC.
The proposal comes amid a broader debate about AI safety. The supplied item says a post by Jacob Coxon nearly two weeks earlier, arguing that Anthropic and OpenAI were taking unacceptable risks, helped trigger international discussion.
The excerpt does not provide the full text of OpenAI’s standards proposal or describe how any global framework would be negotiated or enforced. The report therefore points to an emerging policy initiative rather than a completed international agreement.
Sources
EGazette summarizes reporting from multiple sources; follow the links for the originals.
Related articles
AI Leaders Call for Slower Development, Tighter Oversight
OpenAI reportedly shelved plans for a public listing, saying the timing was not right, according to Deutsche Welle Arabic.
British Columbia sues OpenAI and Sam Altman over Tumbler Ridge school shooting
The Canadian province alleges the deadly February attack could have been prevented had OpenAI alerted police to the shooter's ChatGPT use.
Canada: British Columbia sues OpenAI over ChatGPT's alleged role in Tumbler Ridge school shooting
The Canadian province alleges OpenAI ignored warning signs from a user's conversations with ChatGPT and is demanding changes to the chatbot's safety practices.
Canadian province sues OpenAI after Tumbler Ridge shooting
British Columbia is accusing the California-based company of failing to alert law enforcement to threats made on ChatGPT before a shooting that left nine people dead.
OpenAI creates math advisory group as company reports AI progress on open problems
The company says its systems have resolved more than 100 open mathematical problems, while the new group will advise without directing research.
OpenAI’s George Osborne states datacentre nimbys holding back Britain
A report from The Guardian — World sets out the main available details, with claims kept attributed to their sources.
Comments
Loading comments…