Microsoft says Copilot is for ‘entertainment purposes only’ in its terms

Warnings about trusting AI outputs aren’t just coming from critics — the companies building these tools are issuing similar cautions themselves in their official policies.

Take Microsoft, which is currently focused on expanding the adoption of its AI assistant, Microsoft Copilot, among enterprise customers. At the same time, the company has recently drawn attention on social media over the wording in Copilot’s terms of use, which appear to have been last updated on October 24, 2025.

Within those terms, Microsoft explicitly states: “Copilot is for entertainment purposes only.” The company further warns that the tool can produce incorrect results and may not always function as expected. Users are advised not to depend on Copilot for important decisions or guidance and to use the system at their own risk.

Following the renewed scrutiny, a Microsoft spokesperson told PCMag that the company plans to revise what it described as “legacy language” in the terms. According to the spokesperson, the current wording no longer accurately reflects how Copilot is being used today, and updates will be introduced in a future revision.

The use of such disclaimers is not unique to Microsoft. As noted by Tom’s Hardware, other AI developers have adopted similar approaches. OpenAI cautions users against treating its outputs as a single authoritative source of truth, while xAI similarly advises that its systems should not be relied upon as definitive or factual.

These disclaimers highlight a broader industry acknowledgement that while AI tools are becoming more powerful and widely used, they still require human judgment and verification, particularly when accuracy and reliability are critical.