Exploring Gen's AI Breakthrough for Deepfake Detection
Cyber safety company Gen – the parent company of Norton – has unveiled an early prototype of its latest AI-powered deepfake detection technology, representing a potential shift in how machine learning could combat AI-enabled fraud.
Built in partnership with Intel, the new AI capability can identify simultaneous audio and visual manipulation directly on devices, enabling faster and more private protection without transmitting data to cloud servers.
The technology, first demonstrated at CES 2026 in Las Vegas, leverages on-device AI processing to analyse content in real time as users consume it.
The announcement arrives alongside new research from Gen that could challenge conventional wisdom about deepfake scam distribution patterns. Rather than proliferating through short viral clips, these AI-generated scams appear predominantly during extended viewing sessions where they blend into legitimate content and build persuasion gradually over time.
AI-enabled scams target long-form content
"The presence of a deepfake alone is not the risk," explains Vincent Pilette, CEO of Gen. "Risk emerges when deepfake capabilities are paired with intent: urgent financial requests, promises of guaranteed returns, pressure to act quickly, or instructions to move conversations or payments off platform."
Gen's data indicates that YouTube accounts for the largest share of intercepted deepfake-enabled scam activity, followed by Facebook and X – platforms that dominate long-form, recommendation-driven viewing on TVs and PCs.
Gen's research reveals that most deepfake scam videos are detected during playback rather than as downloads, links or attachments.
"They hide in plain sight, embedded in ordinary video consumption," states Gen. This represents a departure from traditional phishing tactics, where malicious content surfaces through obvious channels like suspicious emails or texts.
The shift towards long-form content exploitation reflects how scammers are adapting their strategies to match modern viewing habits. By embedding fraudulent content within extended videos, they create opportunities to build trust and credibility with viewers over time, making their eventual scam requests appear more legitimate and harder to resist.
Audio AI drives deception tactics
Gen suggests that audio-led deception powered by AI represents a dominant force in the threat landscape.
Many scams rely on cloned or synthetic voices generated by AI paired with slightly modified visuals from legitimate video content. This makes detection particularly challenging for traditional security systems, as the visual component may appear entirely authentic while the audio carries the fraudulent message.
"The tools AI creators rely on for voice cloning, synthetic narration, automated editing and composite visuals are now standard capabilities across many popular platforms and production workflows," Vincent goes on. "Adobe's latest global survey found that 86% of creators use generative AI somewhere in their process."
According to Gen, financial lures remain the primary scam category, with investment advice, trading schemes, cryptocurrency offers and giveaways accounting for the dominant themes. However, the company emphasises that AI-generated deepfakes themselves are not inherently malicious – the risk emerges from how the technology is weaponised.
The accessibility of AI voice cloning tools has lowered the barrier to entry for scammers significantly. What once required sophisticated technical knowledge and expensive equipment can now be accomplished with readily-available software, enabling fraudsters to create convincing impersonations at scale and target victims with unprecedented precision.
AI detection AI at scale
Gen's AI-powered deepfake protection has evolved rapidly since the company introduced on-device detection of AI-generated audio in 2025 through its Norton product line.
Originally launched on AI PCs with Intel and Qualcomm, the capability has now expanded to standard PCs and, at CES, reached a new milestone.
Using Intel's upcoming Panther Lake processor and Gen's image analysis tool, the AI system can detect manipulated videos of public figures directly on the device – which Vincent calls "a new benchmark for the industry".
He adds: "Over time, our technology is expected to expand to detect more than just celebrities and famous people – it will also help to protect against more devastating scams such as family member impersonation. Criminals do not need a million views. They need the right viewer at the right moment."
By prioritising long-form video analysis and in-playback detection using on-device artificial intelligence, Gen aims to identify scams as they unfold rather than after the fact.
