Anthropic and OpenAI urge firms to set AI audit standards
Anthropic and OpenAI are calling for new rules on advanced AI, saying independent testing is needed to keep things safe.
But here's the twist: they're suggesting companies like themselves should set the standards for these audits, not governments.
Some experts worry this could distract from big issues like AI's impact on the environment or its use in the military.
Critics warn self-regulation could hurt startups
Letting tech companies create their own safety checks has raised eyebrows.
Conrad Stosz, a former US official, put it simply: "Lots of evaluators are interested in embedding with labs and getting greater access, but it's a little ambiguous what embedded evaluators means," said Stosz, now head of governance at Transluce.
While Anthropic and OpenAI hope self-regulation will boost trust and help position them as the safest bet for investors, critics say it might make things tougher for smaller startups trying to break in.