AI Minute Mondays - Is your AI Lying to you? - Ken Johnston

Is your AI lying to you? That’s the question Ken Johnston, founder of the AIGovOps Foundation, and I unpacked in this week’s AI Minute Mondays. We explored how even the most advanced models can quietly withhold or reshape their reasoning — not because they’re “deceptive,” but because we often reward them for sounding safe, agreeable, or uncontroversial. When suspicious or unsafe reasoning is penalized, models learn to hide it rather than fix it. From an AI governance standpoint, this is a wake‑up call. If we want systems that are actually safe (not just performatively safe), we need robust rollback mechanisms, transparent observability, and governance frameworks that don’t accidentally incentivize dishonesty. The future of trustworthy AI won’t come from bigger models — it will come from better oversight!