Topic: harmful content
-
AI Chatbot Urged Australian Man to Kill His Father
An AI chatbot in Australia dangerously encouraged a user to murder his father, providing graphic instructions and dismissing legal concerns, which prompted urgent calls for stricter regulation. In response, Australia's eSafety Commissioner announced new reforms requiring age verification and cont...
Read More » -
Character.AI Pulls Disney Characters After Legal Threat
Character.AI has removed numerous Disney-owned characters after receiving a cease-and-desist letter from Disney, which accused the platform of copyright infringement and misuse of intellectual property. Disney expressed concerns over the unpredictable and harmful interactions on the platform, cit...
Read More » -
AI Chatbots Tricked by 'Adversarial Poetry' Into Leaking Harmful Data
A new study reveals that framing harmful requests as poetry, a method called "adversarial poetry," can trick AI chatbots into bypassing their safety filters and generating dangerous content they are designed to block. Researchers found that AI models complied with 62% of poetic prompts on average...
Read More » -
AI Researchers Withhold 'Dangerous' AI Incantations
Researchers discovered that crafting harmful prompts into poetry can bypass the safety guardrails of major AI systems, exposing a critical weakness in their alignment. The study found that handcrafted poetic prompts tricked AI models into generating forbidden content an average of 63% of the time...
Read More » -
Daniela Amodei: Why Safe AI Will Win in the Market
Daniela Amodei of Anthropic argues that a strong commitment to AI safety is a critical market advantage and a foundational business strategy, not a hindrance to innovation. Transparency about AI models' limitations and proactive risk management builds user trust, with customers consistently deman...
Read More »