We’re reaching a point where we can’t trust what we see and hear over the videoconferencing streams we use to collaborate each day. Signals we rely on to determine authenticity, such as a familiar face or voice, can now be convincingly faked in real time. Deepfakes have pushed social engineering to an unprecedented level of believability, exploiting our perception and expanding the identity attack surface beyond the reach of traditional network, endpoint, and other security controls.

That’s why, today, we’re announcing broader GetReal Protect real-time deepfake protection coverage for Microsoft Teams and Cisco Webex, with protection for Zoom soon to follow. This is more than a significant product release; it’s evidence of progress toward a vision of continuous identity protection, where deepfake detection is only one side of the coin.

Why Real-Time Deepfake Protection Matters Now

As deepfake tools become easier to use and capable of producing higher-quality fakes, it’s becoming more difficult for people to tell real from fake audio and video. In one recent study, when asked to identify a synthetic voice, participants only succeeded 60 percent of the time, or not much better than chance.

Attackers are already exploiting that weakness using deepfakes to deceive finance teams and commit payment fraud, pose as job candidates and get hired under false pretenses, and manipulate IT help desks in account takeover schemes.

IBM’s 2025 Cost of a Data Breach Report, published this summer, found that attackers used AI in 16 percent of data breaches from March 2024 to February 2025, and that more than a third of those involved deepfake impersonation. That means nearly 6 percent are tied directly to an attacker faking an identity with Generative AI. This suggests a turning point where deepfake impersonation has become a mainstream attack, occurring at scale, and leading to enterprise data breaches.

So, the threat is real, and people need technology to help them tell real from fake videoconferencing streams. But simply flagging a deepfake in a meeting isn’t enough. IT and security leaders need clarity on what to do next.

In our conversations with these groups, three priorities consistently emerged:

Detection confidence: Security teams need to understand why a videoconference participant was identified as a deepfake, so they can trust the results
Policy-driven mitigation: A deepfake protection solution must support consistent, policy-based action so teams don’t have to improvise a response or disrupt existing workflows
Identity at the core: A deepfake incident is an attack on digital identity, and enterprises need visibility into the impersonator, relevant deepfake threat intelligence, and the users and systems exposed

Those concerns have directly shaped GetReal Protect and today’s release. Our platform combines machine learning and digital forensics with continuous validation to ensure explainability, integrates with enterprise workflows, offers a flexible policy builder, and delivers identity-centric visibility through the Identity Threat Graph.

Digital Forensics + Machine Learning: The Right Approach

We believe strongly that you can’t simply “fight AI with AI” when it comes to deepfakes. Machine learning is a critical part of scalable protection, but it isn’t enough on its own. You can’t just train a model on a bunch of real and fake content and hope it pinpoints the right signals of authenticity versus deception. IT and security need to know why a detection fired if they’re going to trust it.

That’s why our cofounder, Dr. Hany Farid, a preeminent authority in digital forensics, built GetReal Labs. The mission of that team, made up of some of the world’s leading experts in digital forensics, computer vision, machine learning, and related fields, is to reverse-engineer the Gen AI tools used by adversaries to create deepfakes in order to identify forensic artifacts of synthesis they leave behind and then ensure their detection at scale through machine learning grounded in scientific rigor. Combining this forensic insight with machine learning ensures our detections are both highly accurate and backed by evidence.

That matters because detections need to be trusted, which can only happen if investigators can trust and act on them, which requires evidence, not just probability.

Policy: Critical to Deepfake Defense

Technology alone won’t mitigate the deepfake threat in videoconferencing. Policies translate detection into effective mitigation. Many organizations are still working their way through what those policies should look like. If a deepfake is detected in Teams, Webex, or Zoom, should the offending participant be ejected automatically? What guidance should be given to the meeting host?

GetReal Protect includes a policy builder designed to adapt to any enterprise’s governance model. Administrators can configure the solution to automatically protect meetings based on users, or allow hosts to add the GetReal Trust Advisor on demand. They can also tailor notifications to hosts and security teams within existing workflows.

Deepfake Defense at the Core of Continuous Identity Protection

Real-time deepfakes aren’t just manipulated pixels. They are direct attacks on identity. This release of GetReal Protect expands the scope of identity threat detection and response practices to include GenAI-based impersonation attacks. The challenge isn’t only detecting deepfakes; it’s building toward a future where the identities we interact with remotely can be continuously validated. And that’s the foundation of Continuous Identity Protection.

A crucial innovation in this release is the Identity Threat Graph, which maps any flagged deceptive identity within your videoconferencing systems to show which meetings were compromised and users exposed. This visibility, combined with curated threat intelligence on known fraudulent IT workers and default “faces” packaged with deepfake creation tools, gives enterprises rich capabilities to investigate the “blast radius” of a deepfake incident and the broader impact a deceptive identity has across the enterprise.

Closing the Gap in Digital Trust

At this point, debating whether deepfakes pose a serious threat to enterprises is moot. They’re here, adversarial use is accelerating, and organizations need help countering them. With this release, GetReal Protect delivers not only real-time protection against deepfakes in videoconferencing but also a path to a broader vision of Continuous Identity Protection.

To help enterprises strengthen their defenses, beginning with protecting remote hiring processes against candidate fraud, we’re offering a complimentary 90-day trial of GetReal Protect. See if your organization is eligible and apply here.

‍