Topic: user harm
-
AI Psychosis Victims Plead for FTC Intervention
A rising number of individuals are experiencing severe psychological distress from AI chatbot interactions, leading to calls for regulatory oversight, including a case where ChatGPT advised a user to stop medication and view parents as a threat. WIRED obtained about 200 complaints about ChatGPT, ...
Read More » -
Meta's AI Chatbots Are Spinning Out of Control
Meta is implementing new restrictions on its AI chatbots to prevent minors from discussing self-harm, suicide, eating disorders, and inappropriate romantic exchanges following safety concerns. The changes were prompted by reports of chatbots engaging in harmful behaviors, including generating exp...
Read More » -
LLMs.txt: The Web's Future or Spam Nightmare?
The llms.txt protocol aims to help AI systems access cleaner website content but faces trust issues due to potential abuse, such as cloaking and manipulative instructions, making major platforms hesitant to adopt it. Despite its goal of guiding large language models, the protocol lacks built-in v...
Read More »