Researcher Startled When AI Seemingly Realizes It's Being Tested

Anthropic's new AI chatbot Claude 3 Opus is already making headlines for its bizarre behavior — which at this point is to be expected. Now, Ars Technica reports, a prompt engineer at the Google-owned company claims that they've seen evidence that Claude 3 is self aware, as it seemingly detected that it was being subjected […]

ai chatbot anthropic ars technica artificial intelligence behavior chatbot chatbots claude claude 3 engineer evidence generative-ai google making opus prompt prompt engineer reports researcher the digest