Search
Zuckerberg Speaks Out: AI Safety Should Rely on Independent Assessments, Not Collective Slowdowns
1 day ago /
Read about 0 minute
Author：小编
On September 15 (local time), Meta CEO Mark Zuckerberg stated that AI labs should rely on independent assessors and advisors to ensure model safety, without the need for a collective slowdown in development. He pointed out that trust and alignment capabilities are key differentiators for AI models, and those who ignore them will be eliminated. Meta has delayed the release of Muse for several months to focus on safety, but has not required other companies to slow down in tandem. Zuckerberg also mentioned that AI labs face potential legal liabilities, providing an incentive to enhance safety, and that expanding the ecosystem of independent assessments will help improve safety.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
