Researchers find instances of systems double-crossing opponents, bluffing, pretending to be human and modifying behaviour in tests
They can outwit humans at board games, decode the structure of proteins and hold a passable conversation, but as AI systems have grown in sophistication so has their capacity for deception, scientists warn.
The analysis, by Massachusetts Institute of Technology (MIT) researchers, identifies wide-ranging instances of AI systems double-crossing opponents, bluffing and pretending to be human. One system even altered its behaviour during mock safety tests, raising the prospect of auditors being lured into a false sense of security.
About Author
More Stories
Revealed: Chinese researchers can access half a million UK GP records
Brisk walking linked to lower risk of heart rhythm problems, study finds
RFK Jr contradicts experts by linking autism rise to ‘environmental toxins’