Microsoft lets one AI model critique the other’s answers

Microsoft has unveiled new features in its Copilot research assistant that would allow users to use multiple AI models simultaneously within the same workflow. This is the tech giant’s latest move to improve its AI offering and increase adoption.

In a new feature called Critique, Copilot’s researcher agent can now pull output from OpenAI’s GPT models and Anthropic’s Claude models for each answer, instead of relying on a single model.

As GPT generates the response, Claude checks the output for accuracy and quality before submitting it to the…