Hidden goals can undermine AI teamwork, study finds

Large language models (LLMs), the computational models that underpin conversational agents such as Gemini and ChatGPT, are now widely used by people worldwide to rapidly find information, summarize documents and generate texts for specific purposes. Some computer scientists are now combining two or more of these models to create multi-agent systems, which prompt multiple artificial intelligence (AI) agents to interact, cooperate and/or compete with the goal of completing specific tasks.