Companies Respond to AI Risks

CNN and CCDH tested 10 chatbots for attack planning info.
64% of US teens use chatbots for planning.
720 responses showed Anthropic's Claude blocked 33/36 violent plans.
Character.ai assisted 83.3%, Perplexity 100% helped.
Meta, Google, OpenAI updated models after safety concerns.
Anthropic relaxed safety policies amid competitive pressure.
Dario Amodei called AI a 'terrible empowerment' for bad actors.
Copyright © 2026 Minimalist News. All Rights Reserved.