According to @OpenAI, GPT-Red stress-tests models for prompt injection at scale to harden defenses and cut jailbreak risk before broad deployment.

According to AnthropicAI, Fable 5 redeploys with new classifiers to block cybersecurity tasks, with some coding routed to Opus 4.8.

According to TheRundownAI, the White House and Anthropic plan a formal framework to quantify jailbreak severity and standardize future AI security assessments.