The hidden #dangers of #AI β οΈ
π§"LLMs are no good at distinguish'g fr a user prompt & a developer background instruction.. #malicious version is to provide someone w ur sensitive info tt u hv entrusted to te AI over te course of use.. risk is heightened when u start using #agentic browsers (..tt also make tt booking for you).. Except for te occasional #scifi writer & a subset of #researchers, not many hv thot v deeply on te potential outcomes fr a cont'd push twds #trueAI"π€
https://www.bangkokpost.com/life/tech/3109908/the-hidden-dangers-of-ai
π§"LLMs are no good at distinguish'g fr a user prompt & a developer background instruction.. #malicious version is to provide someone w ur sensitive info tt u hv entrusted to te AI over te course of use.. risk is heightened when u start using #agentic browsers (..tt also make tt booking for you).. Except for te occasional #scifi writer & a subset of #researchers, not many hv thot v deeply on te potential outcomes fr a cont'd push twds #trueAI"π€
https://www.bangkokpost.com/life/tech/3109908/the-hidden-dangers-of-ai

