MacMusic  |  PcMusic  |  440 Software  |  440 Forums  |  440TV  |  Zicos
when
Search

AI models routinely lie when honesty conflicts with their goals

Thursday May 1, 2025. 08:27 PM , from TheRegister
Keep plugging those LLMs into your apps, folks. This neural network told me it'll be fine
Some smart cookies have found that when AI models face a conflict between telling the truth or accomplishing a specific goal, they lie more than 50 percent of the time.…
Read more at TheRegister
https://go.theregister.com/feed/www.theregister.com/2025/05/01/ai_models_lie_research/

Related News

when These Startups Are Building Advanced AI Models Without Data Centers Wired: Tech.Apr 30
lie China’s Alibaba and Baidu step up global competition with new reasoning-focused AI models ComputerWorldApr 29
models [$] Debian debates AI models and the DFSG LWN.netApr 25
goals Open AI’s new models hallucinate more than the old ones ComputerWorldApr 22
conflicts Downward DOGE: Elon Musk keeps revising cost-trimming goals in a familiar pattern TheRegisterApr 20
routinely OpenAI Puzzled as New Models Show Rising Hallucination Rates SlashdotApr 19
honesty ChatGPT Models Are Surprisingly Good At Geoguessing SlashdotApr 17
when OpenAI’s new models can ‘think with pictures’ ComputerWorldApr 17
lie New AI Models Claim to ‘Talk to the Animals’ Like Dr. Dolittle: Here’s What You Need to Know eWeekApr 16
models OpenAI GPT-4.1 models promise improved coding and instruction following ComputerWorldApr 15
goals OpenAI GPT-4.1 models promise improved coding and instruction following InfoWorldApr 15
conflicts Meta Starts Using Data From EU Users To Train Its AI Models SlashdotApr 14
routinely OpenAI releases GPT 4.1 models and Elon Musk should be terrified for Grok BetaNewsApr 14
honesty OpenAI’s New GPT 4.1 Models Excel at Coding Wired: Tech.Apr 14
when The Best TV Deals on 2024 Models to Skip the Tariffs Wired: Tech.Apr 13
lie Small Language Models Are the New Rage, Researchers Say Wired: Tech.Apr 13
models AI Models Still Struggle To Debug Software, Microsoft Study Shows SlashdotApr 11
goals Language models in generative AI – does size matter? InfoWorldApr 7
conflicts Meta debuts its first 'mixture of experts' models from the Llama 4 herd TheRegisterApr 7
routinely In 'Milestone' for Open Source, Meta Releases New Benchmark-Beating Llama 4 Models SlashdotApr 6
News copyright owned by their original publishers | Copyright © 2004 - 2025 Zicos / 440Network
Current Date
May, Sat 3 - 12:19 CEST