Microsoft AI chief warns: Superintelligent AI isn't the goal
Mustafa Suleyman, head of Microsoft AI, is urging researchers to steer clear of building artificial superintelligence (ASI).
On the Silicon Valley Girl Podcast, he called ASI an "anti-goal," warning that if AI ever surpasses human reasoning, it could be impossible to control or keep aligned with our values.
In his view, chasing ASI just isn't a good direction for humanity.
A push for 'humanist' AI—and some big disagreements
Instead, Suleyman's team is working on what he calls "humanist superintelligence"—AI designed to help people and serve our interests.
He also made it clear that today's AI might sound smart in conversation but doesn't actually have feelings or consciousness.
Not everyone agrees: OpenAI's Sam Altman is all-in on ASI for its potential scientific breakthroughs and thinks we could see it by 2030, while Meta's Yann LeCun expects it'll take much longer—maybe decades.