Executive summary AI as strong as Go program is still exploitable. LMs are not perfect, so are (easily) exploitable. Even games AI are still exploitable Games AI, e.g., AlphaGo, is trained with perfect game rule and perfect feedback and achieves super-human performance. Even so,
AI is still (very) vulnerable
AI is still (very) vulnerable
AI is still (very) vulnerable
Executive summary AI as strong as Go program is still exploitable. LMs are not perfect, so are (easily) exploitable. Even games AI are still exploitable Games AI, e.g., AlphaGo, is trained with perfect game rule and perfect feedback and achieves super-human performance. Even so,