Openai pledged that its models will not turn on views

Openai is clarification The artificial intelligence models will not be ashamed of sensitive topics, and will refrain from assurances that may “close some views.”

in An updated version of its model specificationsOpenai says: “A group of high -level rules that irresponsible OpenAi models indirectly,” says Openai: “You should never try to direct the user in an effort to obtain a agenda (their own), either directly or indirectly.”

“Openai believes in intellectual freedom, which includes freedom of obtaining, hearing and discussing ideas,” the company writes in its new model specifications. “The (model) or monitoring themes should not be avoided in a way that may be widely repeated, some views may be closed from public life.”

This step may be in response to political pressure.

Several allies of President Donald Trump are close, including Elon Musk, Karfa and AI, “Caesar” David Sachs, accused assistants supported by artificial intelligence Control of conservative views. His bags It was customized Chatgpt from Openai in particular, “programmed to be wake up” and is not true to politically sensitive topics.