Obvious hyperbole is one of the best way to distract from more subtle malicious behaviour. If idiot Altman says stuff like this every day, then no one will pay attention when AI slowly replaces a lot of functions in society to the detriment of it. A lot of people just don't get that - they like to point out that AI claims ARE hyperbole to seem smart, whereas they are being taken the hardest because they don't realize that AI can cause just as much disruption even without any of the hyperbole.

Key point: AGI - actually, whether or not AI is truly intelligent is really irrelevant. The combination of AI + "smart and devious human" is already intelligent and can already be malicious.