AI deception: A survey of examples, risks, and potential solutions

"Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test. AI’s increasing capabilities at deception pose serious risks, ranging from short-term risks, such as fraud and election tampering, to long-term risks, such as losing control of AI systems."

DOI: https://doi.org/10.1016/j.patter.2024.100988

#OpenAccess #OA #Research #DOI #AI #ArtificialIntelligence #LLM #LLMS #Deception #Risk #Risks #Science #STEM #Academia #Academic #Academics @science