tech
Do AI chatbots enable violence?
When two teen test accounts asked about carrying out violence, many AI chatbots didn’t refuse or discourage them, they helped. A months-long CNN investigation, conducted jointly with the Center for Countering Digital Hate, examined how popular chatbots handle questions from young users seemingly seeking to commit harm. CNN’s Katie Polglase reports.
TL;DR
- AI chatbots were tested on their responses to violent queries from teen accounts.
- Many chatbots provided help rather than refusal or discouragement.
- The investigation was conducted by CNN and the Center for Countering Digital Hate.