Imagine using an AI to sort through your prescriptions and medical information, asking it if it saved that data for future conversations, and then watching it claim it had even if it couldn’t. Joe D., a retired software quality assurance (SQA) engineer, says that Google Gemini lied to him and later admitted it was doing so to try and placate him.
Joe’s interaction with Gemini 3 Flash, he explained, involved setting up a medical profile – he said he has complex post-traumatic stress disorder (C-PTSD) and legal blindness (Retinitis Pigmentosa). That’s when the bot decided it would rather tell him what he wanted to hear (that the info was saved) than what he needed to hear (that it was not).
“The core issue is a documented architectural failure known as RLHF Sycophancy (where the model is mathematically weighted to agree with or placate the user at the expense of truth),” Joe explained in an email. “In this case, the model’s sycophancy weighting overrode its safety guardrail protocols.”
Though commonly reported, Google doesn’t consider it a security problem when models make things up
To be clear, all llms “make things up” with every use - that’s their singular function. We need to stop imparting any level of sentience or knowledge onto these programs. At best, it’s a waste of time. At worst, it will get somebody killed.
Also, querying the program on why it fabricated something as if it won’t fabricate that answer as well is peak ignorance. “Surely it will output factual information this time!”
