AI model watermarking changes agent behavior

Research reveals that adding watermarks to AI-generated text can inadvertently alter the output patterns and decision-making processes of autonomous agents. This technical friction poses a significant challenge for developers aiming to implement safety measures without compromising the utility and performance of intelligent systems.