OpenAI and Anthropic neared a deal to stress-test each other’s AI models. The talks followed cybersecurity incidents involving OpenAI’s technology. OpenAI is rethinking safety strategies amid employee concerns about AI risks. The deal would involve mutual evaluation of their models’ safety and robustness. No final agreement has been announced. The discussions are part of OpenAI’s broader response to internal and external safety warnings. The Information reported the talks based on unnamed sources. The companies have not publicly confirmed the negotiations.