Microsoft 365 Copilot Removes Default Alert: “AI Can Make Mistakes”

For users of ChatGPT, a familiar tagline appears at the base of responses: “ChatGPT can make mistakes.” Similarly, Microsoft 365 Copilot displays a disclaimer below its composition box. However, Microsoft has opted to remove this alert by default, citing feedback that it can be distracting for enterprise users. Nevertheless, users who prefer to keep the disclaimer can opt to turn it back on.

According to **Windows Latest**, an update to Microsoft’s Admin Centre confirms that the company will automatically conceal the disclaimer appearing in the Copilot chat interface while using the Microsoft 365 Copilot application.

AI-generated content may be incorrect in Microsoft 365 Copilot

Many users may find themselves puzzled by the naming conventions in Microsoft’s offerings. Here’s a brief overview of the recent terminology changes:

Microsoft 365 (originally the enterprise version of Office) is now labeled Microsoft 365 Copilot.
Microsoft Copilot within the Microsoft 365 app has been rebranded as Microsoft 365 Copilot Chat.
The basic Copilot application is designated for more general everyday tasks.

This alteration specifically pertains to the Copilot chat functionality within Microsoft 365 Copilot.

It’s well understood that AI, despite consistent improvements, is not infallible, and Microsoft aims to empower administrators with greater authority over the visibility of these disclaimers.

The disclaimer reads, “AI-generated content may be incorrect, ” and, with the newest update, it will be hidden by default:

Disclaimer hidden by default in Microsoft 365 Copilot chat

Reasons Behind Microsoft’s Decision to Remove the Disclaimer

Microsoft’s rationale for dropping the alert about AI-generated errors stems from user feedback indicating that it was overly distracting. While the exact number of testers who voiced their concerns remains undisclosed, the company emphasizes that the decision is driven by user input.

In a support document, Microsoft stated, “We’re updating the AI Disclaimer experience in Microsoft Copilot Chat based on customer feedback. Some users found the disclaimer too distracting, while others felt it was too subtle.”

Conversely, there exists a cohort of individuals, particularly within organizations, who find the disclaimer essential for transparency.

For those interested in retaining the disclaimer, Microsoft has introduced a new feature dubbed AI Disclaimer with Heightened Awareness in the Admin center. This feature allows organizations to restore the original disclaimer, improving its visibility without the subtlety of the current iteration.

Enhanced Disclaimer in Microsoft 365 Copilot chat

Enabling this new policy will also render the warning text bolder, making it more noticeable than previous versions. Additionally, it will include a link button, requiring administrators to input a designated URL.

The gradual rollout of the update that eliminates the “AI may be incorrect” alert will commence in the coming weeks, with Microsoft aiming for completion by December 2025.

Source & Images