Using Stable Diffusion (or similar) to get around the new UK face verification requirements

Navigating UK’s New Online Safety Regulations: Can AI-Generated Faces Bypass Age Verification Systems?

In recent developments here in the UK, the government has introduced the Online Safety Act, a comprehensive legislation that enforces stricter content regulation across digital platforms. As part of this initiative, major websites—including social media and adult content platforms—are now implementing age verification procedures to ensure users are legally eligible to access certain content. This has sparked widespread discussion and concern among internet users about privacy, accessibility, and the effectiveness of these new measures.

Many platforms are adopting innovative verification methods, some of which involve real-time facial recognition technology. Users are prompted to display their faces on camera, and AI algorithms assess whether the individual appears to meet the age threshold. While these systems are designed to promote safety, they are not without flaws—particularly given their reliance on facial analysis, which is inherently probabilistic and can often produce inaccuracies.

This raises an intriguing question: could AI itself be harnessed to circumvent these verification processes? Some tech-savvy individuals have experimented with this idea, using AI image-generation tools—such as Stable Diffusion—to create synthetic faces that resemble a specific age demographic, like a man in his 30s.

For example, a group of enthusiasts attempted to generate a convincing adult male face using a selection of pre-trained models. One fortunate participant had access to several models stored locally, especially important given that some AI websites—like Civit AI—are now blocked in the UK, making direct access impossible due to strict legislation. Interestingly, the generated faces appear quite realistic from a frontal perspective, but challenges remain when the AI is asked to simulate head turns or profile views. Producing dynamic, multi-angle videos of artificial faces is still beyond current capabilities, though this could be an area of future development.

Additionally, some users have resorted to VPNs—whose sales are reportedly surging in the UK—to access these verification systems from outside the country. By setting their virtual location to a different country where such restrictions are not in place, users can test whether these AI-generated faces can pass the age checks. Many adult websites now incorporate facial verification, so experimenting with VPNs and AI-generated visuals may shed light on the robustness of these systems.

Of course, while using a VPN is an option, it isn’t always appealing or practical for everyone. Many individuals prefer to avoid additional costs associated with VPN subscriptions unless necessary. Moreover, a significant number of adult sites still do not enforce strict verification procedures,