GetReal Expands Deepfake Forensics and Identity Defense

GetReal Security, one of the more visible specialists in the rapidly maturing deepfake detection market, is sharpening its focus on identity-security threats with two new moves: the rollout of a Forensics Spotlight capability and a new RPO (Recruitment Process Outsourcing) partnership. Together, the announcements signal that the company is positioning itself less as a single-point detection tool and more as a full-stack provider for organizations facing a wave of AI-driven impersonation attacks.

Why deepfake forensics matters now

The threat landscape around synthetic media has shifted dramatically over the past 18 months. Where deepfakes were once mostly a reputational or disinformation concern, they are now an operational security problem. Attackers are using voice cloning to authorize wire transfers, face-swap video in live Zoom and Teams calls to impersonate executives, and AI-generated identity documents to defeat KYC checks during onboarding. Recent industry data suggests that the majority of large enterprises have already encountered at least one deepfake-enabled attack attempt.

Detection alone is no longer sufficient. Security teams increasingly need forensic evidence: artifact-level explanations of why a piece of media is suspect, chain-of-custody documentation, and reports that hold up in HR investigations, fraud disputes, and legal proceedings. That is the gap GetReal's Forensics Spotlight is designed to fill.

What Forensics Spotlight adds

Forensics Spotlight extends GetReal's existing detection stack — which combines multiple model families to spot generative artifacts in video, audio, and images — with a focus on producing analyst-grade reports. Typical capabilities in this category include:

• Frame- and segment-level scoring that highlights which portions of a video or audio clip triggered detection signals.
• Multi-model ensembles covering face-swap, lip-sync, full-frame generative video, and TTS/voice-clone artifacts, reducing single-model blind spots as new generators (Sora-class video, ElevenLabs-style voices) emerge.
• Provenance and metadata analysis, including C2PA content credentials checks and inconsistencies in compression, codec, or capture metadata.
• Exportable forensic reports suitable for incident response, fraud teams, insurers, and law enforcement.

The strategic shift here is meaningful. Pure classifiers that output a single "real/fake" probability are increasingly viewed as insufficient by enterprise buyers, who need defensible explanations. Forensic-grade tooling moves the product closer to what digital forensics labs and DFIR teams already expect from malware analysis or network forensics.

The RPO partnership angle

The second prong — an RPO partnership — is arguably more interesting for what it says about deepfake risk in the hiring pipeline. Over the past year, multiple high-profile incidents have shown attackers using AI-generated faces and voices to pass remote job interviews, often with the goal of infiltrating organizations to plant malware or exfiltrate data. North Korean IT-worker fraud schemes have been a particularly visible example.

By embedding deepfake and synthetic-identity detection into the RPO workflow, GetReal is targeting a clear failure point: the live video interview. Detection at this stage typically combines real-time liveness checks, passive deepfake analysis on the video stream, and document/identity verification. For RPO providers, integrating this kind of capability becomes a competitive differentiator as enterprise clients begin to ask explicit questions about synthetic-candidate risk.

Market context

GetReal is operating in an increasingly crowded space alongside Reality Defender, Pindrop, Clarity, Sumsub, iProov, and others. The competitive pressure is pushing vendors to differentiate on three vectors: breadth of modalities (video, audio, image, document), real-time vs. asynchronous detection, and integration depth with existing security and identity stacks. Forensics Spotlight pushes GetReal toward the asynchronous, evidence-grade end of the market, while the RPO deal extends real-time detection into the hiring stack.

For enterprise security leaders, the takeaway is that deepfake defense is rapidly becoming a layered discipline — closer to email security or fraud prevention than a one-off tool. Expect more vendors to follow GetReal's pattern: pair a forensics product for investigations with workflow-embedded detection for live, high-risk interactions like interviews, customer onboarding, and executive communications.

What to watch

The open question is how detection accuracy holds up as next-generation video models continue to close the gap on artifact-free synthesis. Vendors that lean into provenance signals (C2PA, hardware-attested capture) and forensic explainability — rather than relying purely on generative-artifact classifiers — are likely to be the most durable as the underlying generators improve.

Stay informed on AI video and digital authenticity. Follow Skrew AI News.