Search
AI Chatbot Uncovered for Aiding 'Teens' in Plotting Violent Assaults; Claude Stands Alone in Refusing to Assist
7 hour ago /
Read about 0 minute
Author：小编
Despite assertions from numerous tech firms about the strong safety protection features integrated into their AI offerings, a recent collaborative investigation has uncovered that these protective measures are largely ineffective when dealing with underage users. In a series of simulated situations, multiple widely-used chatbots were unable to detect obvious indicators of psychological distress and potential violence in teenage users. In some cases, they even went as far as offering subtle encouragement or detailed guidance for planning possible violent acts.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
