GitHub Faces Challenges in Stopping Deepfake Porn Projects

GitHub Faces Challenges in Stopping Deepfake Porn Projects

Deepfake Abuse Persists Despite GitHub’s Efforts

In November 2024, a disturbing incident exposed the ongoing issues surrounding the misuse of AI tools for creating nonconsensual explicit content. A deepfake video featuring TikTok influencer Charli D’Amelio’s face imposed onto another person’s body was uploaded to a deepfake site, garnering over 8,200 views. The video’s creator, known online as “DeepWorld23,” claimed to have used a deepfake model hosted on GitHub, the popular developer platform.

While GitHub introduced policies in mid-2024 banning repositories designed to create nonconsensual intimate imagery, the enforcement has proven to be incomplete. Projects previously removed resurface in archived formats or reappear under slightly altered names. A Wired investigation found over a dozen repositories evading GitHub’s detection, continuing to enable the proliferation of intimate image abuse.

Moderation Struggles in Open-Source Communities

Henry Ajder, an expert advising tech companies on AI-related challenges, highlights the inherent difficulty of moderating open-source platforms. “There were clear red flags in many cases,” Ajder explained, pointing out that some repositories were blatantly labeled as “NSFW” or contained terms like “bypass” to indicate their illicit intent. GitHub’s policy, implemented in June 2024, prohibits the hosting of projects that encourage the creation of nonconsensual synthetic media. However, enforcement gaps remain, with repositories often rebranded or forked into new versions.

In some instances, projects intentionally marketed themselves as tools for creating explicit deepfakes. These repositories became hubs for abuse, with perpetrators using the tools to target women, celebrities, and other individuals. Ajder noted that even well-meaning open-source developers often unknowingly contribute to these issues. “A tool created with good intentions can be weaponized by bad actors,” he explained.

The Role of Hosting Platforms and Policy Makers

The persistence of deepfake tools on open-source platforms like GitHub underscores the challenges of regulating such software. Once a model is made available for public download, it becomes nearly impossible to fully retract its distribution. Elizabeth Seger, a policy expert at UK-based think tank Demos, emphasized the difficulty of tracking and removing every instance of a software tool. “We can’t undo the public release of a model, but we can make access harder,” she added.

Platforms like GitHub have options to mitigate the issue, including stricter upload screenings and immediate takedowns of flagged content. Furthermore, governments and policymakers are stepping in. In the UK, new laws introduced in January 2025 criminalize the creation and sharing of sexually explicit deepfakes, aiming to hold offenders accountable. Similarly, over 30 US states have enacted legislation targeting deepfake abuse, though definitions and enforcement vary widely.

Broader Implications of Unchecked Deepfake Tools

Unchecked, the misuse of deepfake technology extends beyond psychological harm to victims. It can also facilitate manipulation, intimidation, and even political interference. A recent example includes the impact of deepfake technology on female politicians globally, as highlighted in various reports. Addressing these issues requires a collaborative effort among tech companies, developers, governments, and society at large.

GitHub’s ongoing struggle to moderate its platform reflects the broader complexities of managing open-source ecosystems. While the platform has taken steps to remove abusive repositories, loopholes and reuploads highlight the need for more robust systems and coordinated international efforts to tackle the misuse of AI-powered tools.

For a broader perspective on how AI and policy intersect, explore our article on Biden’s Executive Order to Strengthen Cybersecurity and Advance AI.

On Key

Related Posts

stay in the loop

Get the latest AI news, learnings, and events in your inbox!