In This Article
Microsoft is preparing a Teams feature designed to help organizations spot deepfake impostors during meetings. Rather than build its own detection system, Microsoft plans to let certified third-party providers analyze audio and video for signs of manipulation and send alerts into Teams, where meeting organizers can decide how to respond.
Key takeaways
The planned safeguard adds specialist analysis to Teams, but Microsoft has not yet identified the providers that will support it.
- Certified third parties will assess meeting audio and video for possible manipulation.
- Teams will surface controls for organizers to respond to suspicious signals.
- AI-generated impersonation is already a concern in hiring and financial scams.
Why deepfake impersonation matters to businesses
Criminals have used synthetic or altered video and audio to pose as job applicants and exploit trust in familiar colleagues. The FBI warned about deepfake-assisted applicant impersonation in 2022 and renewed its warning in 2025, citing North Korean IT workers using AI and face-swapping tools to conceal their identities during interviews.
The risk also extends to financial and information-security scams conducted over video calls. A Gartner survey of 3,000 job seekers found that 6% said they had impersonated someone in an interview or had another person take their place. Gartner has projected that one in four candidate profiles worldwide could contain false identity information by 2028.
How the Teams safeguard is expected to work
Microsoft says certified third-party services will analyze meeting audio and video for signs of synthetic or manipulated content. Those services will pass signals to Teams, which will give organizers controls to act when a meeting appears suspicious. Microsoft has not named participating providers, and the available details do not establish that the feature can conclusively identify every deepfake.
Detection can be difficult because video compression, changes in lighting, blur, background movement and poor audio may obscure clues. That makes the feature a potential additional signal for meeting hosts—not a substitute for sound identity checks and established approval procedures.
What enterprise leaders should consider
Organizations can reduce exposure by treating unusual requests made in meetings—especially payment, credential or sensitive-data requests—as requiring independent verification. Confirming a high-impact instruction through a known, separate channel can help limit reliance on a voice or face alone. Before adopting third-party analysis, IT and security teams should also review provider capabilities and applicable data-handling requirements.
For data leaders, identity assurance is part of the broader discipline of trustworthy operations: analytics and reporting are only as reliable as the information and processes behind them. Spargent helps enterprises build scalable Microsoft Fabric and Power BI data platforms; those platforms can support clearer, faster decision-making, while meeting identity safeguards address a separate but related source of operational risk.
