YouTube deepfake detection: what changes for brands and SMEs

What has changed: the likeness detection feature is now available to everyone

Starting in May 2026, YouTube has extended its program of AI likeness detection to all adult users of the platform. This is reported by The Verge , who 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.

Furthermore, YouTube specified that the number of removal requests generated so far has been "very small." 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 believable fake videos. Today, you don't need an advanced post-production team. A few minutes of original footage and an AI model accessible even to non-techies are enough.

According to an analysis published by McKinsey , the adoption of generative AI tools in organizations has grown exponentially in the last two years. Consequently, 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 harder to contain.

The immediate impact on video content strategies

For SMEs that use YouTube as a communication channel — both for the 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's designed for individual protection, not for brand protection as a legal entity. Therefore, a company cannot register 'the brand's face' in an abstract sense. It must do so through the physical people who represent it.

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

How YouTube's system works technically

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

Like other AI content moderation systems, this tool is not infallible. YouTube has not made the rate of false positives and false negatives public. Furthermore, the platform has not clarified whether the system can detect 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 at SHM Studio we suggest that SMEs address this issue on three distinct levels, which can be activated in parallel without requiring extraordinary investments.

• Pre-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 measure with high protective value.
• Audit of existing content: it's useful to check if unauthorized content already exists using company faces. A systematic analysis of search results — integrated with tools SEO monitoring — can reveal anomalous presence of cloned content.
• Internal risk management policy: define who has the authority to request content removal, within what timeframe, and through which channels. This process should be documented and integrated into the plan of Digital marketing company.

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 campaigns that use videos with real faces are particularly vulnerable. A convincing deepfake simulating a company's founder while promoting a fake product can generate direct commercial damage, as well as reputational harm.

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 during which false content can circulate and cause damage.

For SMEs investing in advertising video , this topic 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 content abuse.

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 normalize. In other words, companies might delegate responsibility entirely to the platform, reducing their own active vigilance.

However, automated systems have inherent limitations. They cannot assess context, they don't know brand strategy, they don't distinguish between satirical use and attempted fraud. Therefore, human supervision remains essential.

We at SHM Studio we believe that the adoption of these tools should be accompanied by a corporate culture aware of AI risk. This applies to the AI strategy broadly speaking, 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's reasonable to expect that the likeness detection feature will also extend to content generated by business accounts, not just personal profiles. Furthermore, it's likely that platforms like LinkedIn — increasingly central for B2B strategies — introduce similar tools.

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

Finally, it's worth monitoring the evolution of European regulations. The AI Act already provides transparency obligations for synthetic content. Consequently, the regulatory framework could impose stricter standards on platforms that currently act voluntarily. To delve deeper into the operational implications of these developments, you can contact the SHM Studio team or check out the Blog for periodic updates on AI and digital strategy.