An unprotected database belonging to Chinese AI company DeepSeek exposed highly sensitive information, including chat history, secret keys, and backend data.
The only way to advance AI safety is to increase human interactions, human values and societal governance to promote a reinforced human feedback loop,...