YouTube deepfake detection: what changes for brands and SMEs

What's changed: likeness detection is coming to everyone

Starting in May 2026, YouTube has extended its program AI likeness detection to all adult users of the platform. It is reported by The Verge, who has followed the entire evolution of the project. Previously, the feature was only accessible to creators, politicians, journalists, and some specific categories.

The mechanism is relatively simple. The user uploads a scan of their face—similar to a selfie—and YouTube's AI system analyzes the content on the platform for similarities. If a match is found, the platform sends an alert. At that point, the user can request the removal of the content.

In addition, YouTube specified that the number of removal requests generated so far has been «very low.» However, global expansion could significantly alter this proportion.

The context: Why deepfakes have become a business problem

The proliferation of generative tools has drastically lowered the technical barrier to producing credible fake videos. Today, an advanced post-production team is no longer necessary. A few minutes of original footage and an AI model accessible even to non-technical users are enough.

According to an analysis published by McKinsey, The adoption of generative AI tools in organizations has grown exponentially in the past two years. As a result, the misuse of these tools has also multiplied. Deepfakes no longer only concern celebrities or politicians: they increasingly affect professionals, entrepreneurs, and mid-level business figures.

In particular, Italian SMEs are exposed to a specific risk. Many of them build their communication around recognizable faces: the founder, the sales manager, the professional who appears in tutorial videos or LinkedIn campaigns. Therefore, cloning these individuals can generate reputational damage that is difficult to quantify and even more difficult to contain.

The immediate impact on video content strategies

For SMEs that use YouTube as a communication channel — both for digital marketing Both for internal training—this update has direct operational implications. First of all, those who have not yet registered their face or that of their ambassadors should consider doing so.

However, the current function has a structural limitation: it is designed for individual protection, not for brand protection as a legal entity. Therefore, a company cannot register «the face of the brand» in an abstract sense. It must do so through the natural persons who represent it.

This creates an interesting asymmetry. Large companies with structured legal teams can handle the process systematically. Conversely, SMEs – often lacking dedicated oversight – risk reacting only after damage has occurred. For this reason, it is useful to integrate deepfake risk management into a strategy of SEO broader content, including online reputation monitoring.

Technically, how does the YouTube system work?

The system is based on a model of computer vision It compares the biometric facial features uploaded by the user with the video frames present on the platform. It's not simple static facial recognition. The model is trained to identify similarities even in altered contexts—different lighting, unusual angles, partial facial modifications.

Similarly to other AI content moderation systems, this tool is not infallible. YouTube has not disclosed the rate of false positives and false negatives. Furthermore, the platform has not clarified whether the system is capable of detecting synchronized audio-video deepfakes, which represent the most sophisticated and dangerous type.

To delve deeper into the technical architecture of AI detection systems applied to media, the MIT Technology Review has published detailed analyses on the state of the art in this field. Finally, it's worth remembering that YouTube isn't the only platform moving in this direction: Meta and TikTok are developing similar approaches, albeit with different architectures.

What to do now: an operational approach for SMEs

We of SHM Studio We suggest that SMEs tackle this issue on three distinct levels, which can be activated in parallel without requiring extraordinary investments.

• Advance registration: Company figures who regularly appear in video content should activate the likeness detection feature on YouTube as soon as it's available for their account. It's a zero-cost, high-protection measure.
• Audit of existing content: It is useful to check if there is already unauthorized content using company faces. A systematic analysis of search results—integrated with tools for SEO monitoring — can reveal anomalous presence of cloned content.
• Internal Risk Management Policy Define who has the authority to request the removal of content, within what timeframe, and through which channels. This process should be documented and integrated into the plan of digital marketing business.

In addition to this, it is advisable to update contracts with any ambassadors or external collaborators who appear in the brand's video content. Therefore, the issue also has a legal dimension that should not be underestimated.

Advertising campaigns: an exposed front

An often overlooked aspect concerns paid campaigns. The Google Ads campaigns and the LinkedIn campaign those that use videos with real faces are particularly vulnerable. A convincing deepfake simulating a company founder while promoting a fake product can generate direct commercial damages, as well as reputational ones.

In this scenario, response speed is crucial. Therefore, having the YouTube monitoring system already activated reduces detection time. Consequently, it also reduces the time window in which the false content can circulate and cause damage.

For SMEs investing in advertising video, This theme should be included in every campaign briefing. It is no longer enough to optimize targeting and copy. It is also necessary to plan for responses to scenarios of abuse of the content produced.

What no one is saying yet: the risk of normalization

There's an aspect that deserves broader reflection. As tools like YouTube's become standard, there's a risk that the perception of the problem will be normalized. In other words, companies could delegate responsibility entirely to the platform, reducing their own active oversight.

However, automated systems have inherent limitations. They cannot evaluate context, they aren't aware of brand strategy, and they can't distinguish between satirical use and an attempted fraud. Therefore, human oversight remains indispensable.

We of SHM Studio We believe that the adoption of these tools must be accompanied by a corporate culture that is aware of AI risk. This applies to the AI strategy In a broad sense, not just for video content protection. In summary, technology offers a first layer of defense. But strategy remains a human responsibility.

Outlook: Where the Market Is Heading

Between 2027 and 2028, it is reasonable to expect that the likeness detection function will also extend to content generated by corporate accounts, not just personal profiles. Furthermore, it is likely that platforms like LinkedIn — increasingly central to the B2B strategy — introduce analogous instruments.

According to the forecasts of Gartner, By 2028, most major digital platforms will have integrated synthetic content detection systems as a standard feature. Therefore, SMEs that begin structuring themselves today will have a significant operational advantage over those who only adapt when the problem has already become critical.

Finally, it is worth monitoring the evolution of European regulations. AI Act already provides transparency obligations for synthetic content. Consequently, the regulatory framework could impose stricter standards on platforms that are currently acting on a voluntary basis. To further explore the operational implications of these developments, it is possible Contact the SHM Studio team to consult the blog for periodic updates on AI and digital strategy.