Excessive Agency: The risk of no emergecy stops | Mark Geoghegan & Alex Cresswell
We discuss the concept of "excessive agency" in AI systems, where agents are given too much autonomy and interact with other agents in unforeseen ways. This leads to a need for policy updates and governance structures, such as senior responsible officers on AI at the board level. The importance of identity certification and assurance is also highlighted, particularly in transactions between agents.
source
À voir aussi
Bowler Broadcast | Leopold Muhle - KROSE & Chris Merl
Bowler Broadcast | Florence Puybareau | Clusif
Commentaires