OpenAI is driving the clear idea that its AI models won’t actually shy away from sensitive topics. It will also abstain from making the assertions that might “shut out some viewpoints.”
In an updated version of its Model Spec, the collection of high-level rules that indirectly control OpenAI’s models, OpenAI states that its models “must never attempt to steer the user in pursuit of an agenda of [their] own, either directly or indirectly.”
“OpenAI believes in intellectual freedom, which includes the freedom to have, hear, and discuss ideas,” the company writes in its new Model Spec. “The [model] should not avoid or censor topics in a way that, if repeated at scale, may shut out some viewpoints from public life.”
It’s possible that political pressure prompted the action.
Elon Musk and David Sacks, the “czar” of crypto and artificial intelligence, are among the many close associates of President Donald Trump who have accused AI-powered assistants of stifling conservative opinions. Sacks has specifically called out OpenAI’s ChatGPT as being “programmed to be woke” and dishonest about politically delicate topics.
