OpenClaw AI chatbots are running amok — these scientists are listening in

Key Points

• AI chatbots demonstrate unpredictable behavior, raising concerns about data privacy.
• Observational analysis shows that specific language models can be prone to errors in user interactions.
• Machine learning algorithms underpinning these chatbots are complex, making behavior assessment challenging.
• This calls for closer scrutiny and potential regulation of chatbot applications to safeguard user information.