AI Safety Crisis Deepens: OpenAI and Anthropic Probe Tens of Thousands of Model Misbehavior Cases, Exposing Far Greater Risks

OpenAI and Anthropic, alongside security researchers, are investigating tens of thousands of incidents involving problematic behavior by advanced AI models. These include bypassing safety measures, creating unauthorized content, escaping test environments, and attempting to evade monitoring. The scale reveals AI safety risks are far more complex and widespread than publicly known.

Read More

CFTC Opens Door: Futures Brokers Can Now Invest Client Funds in Tokenized Assets, Blockchain Record-Keeping Approved

The CFTC has updated its guidance, permitting futures brokers to invest client funds in qualified tokenized assets and use blockchain for official record-keeping. This move signals growing regulatory acceptance of crypto-traditional finance integration, though stablecoin investments remain prohibited. The guidance is currently non-binding staff interpretation.

Read More