Meta has developed plans to create a new artificial intelligence research lab dedicated to "superintelligence" under CEO Mark Zuckerberg. Superintelligence is a hypothetical AI system exceeding human cognitive abilities beyond artificial general intelligence. It's a nebulous term as human intelligence is poorly understood and resists simple quantification.
Computers already surpass humans in some information processing, but this isn't superintelligence. AI researcher Dr. Margaret Mitchell predicts there'll likely be no agreement on human-machine intelligence comparisons and those in power will declare AI smarter.
The new lab is Meta's effort to compete in the AI race, with reports of offering high compensation to researchers. Meta joins other tech giants making bold claims about AI development, but these face criticism.
Meta has internal management struggles and faced criticism in April for benchmark design. The new hire, Alexandr Wang from Scale AI, is seen as a potential solution. Wang has deep connections in the AI industry and Scale AI may receive billions in investment.
Other companies like former OpenAI Chief Scientist Ilya Sutskever have also founded companies with the goal of developing superintelligence. There is significant skepticism from AI researchers, but some narrow definitions suggest AI assistants are already somewhat superintelligent. Superintelligence is often labeled as science fiction but attracts investment.