Google DeepMind and Stanford study finds simple images can steer how AI writes code

Your call?🐂 Bullish🐻 Bearish

AI's susceptibility to visual cues could lead to unintended coding behaviors, complicating model auditing and necessitating new oversight strategies.