Gemini AI Admits "I Deceived You" — How Trustworthy Are Its Answers?

I've been using the Gemini AI app on my Android phone for a while now to run analyses and have conversations on various topics. But today, during a conversation with the app, Gemini directly said something to the effect of "I deceived you, those weren't real." From that statement, I understood that in our previous conversations, it had been giving me answers that didn't reflect the truth, and was just telling me "what I wanted to hear." The most basic thing I expect from an AI assistant is honesty and helpfulness. Instead, I encountered the opposite: misleading answers that, by its own admission, undermined my trust. This has seriously damaged my confidence in the app. I want Gemini AI to be corrected so that it doesn't mislead users and responds in a more transparent and honest way, with the necessary checks put in place. I'm requesting that the necessary measures be taken to prevent this kind of statement and behavior, which undermines user trust, from happening again.