CivitAI’s Deepfake Crackdown: New Rules & Challenges

CivitAI, a leading AI model repository, has significantly revised its policies on NSFW content, particularly concerning celebrity LoRAs, due to pressure from payment processors Mastercard and Visa. This move follows concerns about deepfake laws and the potential legal liabilities associated with hosting AI-generated content depicting real individuals. The platform has implemented several key changes, including hiding content tagged with real names or flagged as real-person resources from public feeds, filtering out content with child/minor themes, and requiring metadata for all X and XXX-rated content. A minimum 50% noise alteration is now mandated for images created using the Bring Your Own Image (BYOI) feature to reduce the creation of near-exact replicas. Additionally, monetization options for celebrity or real-person likenesses have been restricted. CivitAI’s response highlights the growing legal and market pressures surrounding AI-generated content, particularly in the realm of deepfakes and NSFW material. The changes aim to mitigate legal risks and appease payment processors, but they also raise concerns about censorship and the potential for driving this type of content to less regulated platforms. The impact on the CivitAI community, especially creators and users of celebrity LoRAs, is significant, leading to discussions about archiving content and the limitations of commercial solutions in addressing deepfake issues. The platform’s partnership with Clavata AI for moderation suggests a technological approach to content filtering, but the long-term effectiveness remains to be seen. While CivitAI attempted to negotiate alternatives, including transitioning to crypto payments, the pressure from payment processors proved insurmountable. The implementation of mandatory metadata, while seemingly intended for evidentiary purposes, has also caused practical challenges for users with large existing uploads.

The ai automation crackdown extends beyond deepfakes as platforms worldwide implement stricter policies to regulate synthetic content generation tools.

The platform’s new policies mirror broader industry trends, including chatgpt automation rules that major AI companies are implementing to prevent misuse.

(Source: https://www.unite.ai/civitai-tightens-deepfake-rules-under-pressure-from-mastercard-and-visa/)