Sam Altman: OpenAI Is Not Secretly Sitting on a Conscious AI Model

Sam Altman says OpenAI does not secretly have a conscious model or one capable of self-improvement today, adding that the bigger risks come from how increasingly powerful models could be misused.
He sees potential risks around bioterror, cybersecurity, and other forms of misuse, and believes future self-improving models could create a genuine loss-of-control risk.