✉news SciencePhysics first seen 6 d ago, last 6 d ago, peak #33
Physicists propose formula to detect undesirable AI responses
Original: Physicists propose formula for detecting undesirable AI responses — Live Science
Physicists have proposed a formula aimed at identifying undesirable responses from artificial intelligence systems, as reported by Live Science. The work applies physics-style quantitative methods to the problem of flagging AI outputs that are misleading, harmful or otherwise unwanted. The proposal is drawing attention as developers and regulators search for practical tools to measure and control AI behaviour.
Why now: Concern about unreliable or harmful AI outputs is high, making any quantitative detection method newsworthy.
Live Sciencephysicistsartificial intelligence
Evidence
API: https://socialmediatrends-api.osmike.com/v1/trends/2306024