Leading AI labs like Anthropic and OpenAI conduct extensive red teaming, with OpenAI utilizing external security researchers for GPT-4 before its public release to refine safety systems.
Open the full topic