Avast launches Deepfake Guard & expands Scam Guardian

Avast has launched a new Windows PC feature called Deepfake Guard and expanded the availability of its Scam Guardian tools on mobile devices worldwide.

The company said Deepfake Guard uses artificial intelligence to analyse audio in video content and flag suspected scam deepfakes. It sits alongside Avast Scam Guardian and Scam Guardian Pro on mobile, which focus on scam calls, texts and emails.

Avast said the combined release broadens its scam protection across PCs and smartphones. The company positioned the move as a response to the growing use of manipulated media in fraud and the spread of scams inside mainstream content.

Deepfake detection

Deepfake Guard runs on Windows PCs and forms part of Avast Premium Security. Avast said it supports English-language analysis across Facebook, DailyMotion, Instagram, TikTok, Twitch, Vimeo, X and YouTube.

The company said the feature works on-device rather than in the cloud. It said this approach improves speed and privacy. Avast also said the tool warns users before they become victims.

Avast described different operating modes depending on the PC. It said the feature works best on newer "AI PCs" built on Intel and Qualcomm chipsets. It also said it can run on traditional PCs with manual detection or with automated detection enabled in some cases.

On lower specification devices, Avast said Deepfake Guard is an opt-in download. It said users can choose to install it on Windows 11 PCs with 8GB of RAM and four CPU cores. On higher-end traditional PCs, Avast said the system requires Windows 11 and at least 16GB of RAM and six or more CPU cores.

Mobile expansion

Avast said Scam Guardian and Scam Guardian Pro now have full international availability on mobile devices. The company positioned the mobile tools as part of the same scam protection suite as Deepfake Guard for Windows.

The firm did not set out a breakdown of features between Scam Guardian and Scam Guardian Pro in its announcement. It described coverage across scam calls, texts and emails.

Rising video scams

Gen Threat Labs, the threat intelligence unit of Avast owner Gen, reported a rise in deepfake-driven fraud attempts. The group said it recorded 159,378 instances of unique deepfake scams in the fourth quarter of 2025 on devices where its video scam detection was active.

Gen said these cases combined manipulated media with what it described as clear scam intent. The company framed the figure as evidence of how frequently scammers use such techniques.

In New Zealand, Avast reported a sharp increase in "fake tutorial" scams. It said this category surged by 152% in the fourth quarter.

The company described fake tutorial scams as "how to" videos on YouTube and similar step-by-step guides on other sites. It said these videos claim to offer free downloads or software. It said the videos often direct viewers to download fake guides or malicious applications.

Platforms in focus

Avast said YouTube accounted for the largest share of blocked deepfake-enabled scam videos on PCs. It said Facebook ranked second, followed by X.

The company said many deepfake scams appear as part of normal viewing. It said they often do not arrive as downloads, attachments or links. It said this makes them harder to spot because they blend into everyday video consumption.

"Deepfakes are a serious threat, but they're a symptom of a larger problem: deception," said Leena Elias, Chief Product Officer at Gen.

"AI-generated content on its own isn't inherently harmful. The risk comes when scammers use it to create urgency, apply pressure, and take advantage of trust. That's what Avast Deepfake Guard is built for: helping people understand when something may be manipulated and make safer decisions in the moment. By expanding Avast Scam Guardian globally on mobile, we're empowering people worldwide to identify and avoid scam calls, texts and emails with confidence," said Elias.

Avast said it will continue to develop the scam protection suite across Windows PCs and mobile devices as scam tactics evolve and video-led fraud becomes more common across mainstream platforms.