Piracy Lawsuit Targeting Meta Challenges Accountability in AI Creation

Meta Confronts Lawsuit Over Alleged Use of Pirated Adult Content for AI Training

Meta Platforms Inc. is attempting to have a lawsuit dismissed that was brought by Strike 3 Holdings LLC, an adult film distributor based in Miami. The suit accuses Meta of unlawfully downloading thousands of pornographic videos to train its artificial intelligence systems. Meta has dismissed these claims as "illogical and baseless," asserting that the accusations rely on "speculation and insinuation" rather than any solid proof of AI training, according to

. The complaint, which was filed in July, alleges that used both company and concealed IP addresses to torrent close to 2,400 adult films since 2018 as part of efforts to build advanced multimodal AI, as detailed by Ars Technica.

According to Strike 3, Meta’s alleged downloads—including those routed through a “stealth network” of 2,500 untraceable IPs—were part of a coordinated plan to create a secret adult-oriented version of its AI video generation tool, as Ars Technica reported. In its motion to dismiss, Meta argues that the supposed activity—an average of just 22 downloads per year across 47 IP addresses—does not match the scale or intent needed for systematic AI training, as also noted by Ars Technica. Meta further points out that its terms of service strictly ban the creation of adult content, making the lawsuit’s claims unlikely, according to Ars Technica.

Meta also maintains that these downloads were probably for personal use by staff, contractors, or external parties, not for any company AI project. The company’s filing notes that many of the downloads happened before Meta’s major AI research began, and there is no direct evidence linking the activity to specific employees or AI models, as Ars Technica observed. The filing states, “Monitoring every file downloaded by anyone using Meta’s global infrastructure would be an extremely difficult and intrusive task,” rejecting the idea that Meta should be liable for unverified torrenting by unidentified individuals, as Ars Technica also reported.

This lawsuit is part of a larger trend of copyright cases aimed at AI companies. In a related matter, a federal judge in New York recently rejected OpenAI’s attempt to dismiss a lawsuit from authors who allege their works were used to train ChatGPT, according to

. The decision allows claims that AI-generated content infringes on copyrights to move forward, adding complexity to the legal environment for technology firms developing large language models.

Meta’s financial reports also show the increasing costs associated with AI development. The company’s third-quarter results were mixed, with operating margins dropping to 40% from 43% in 2024, due to a 28% increase in research and development spending, according to

. Meta has repeatedly raised its projected AI-related spending for 2025, now estimating $71 billion in expenses, and expects capital investments to rise further in 2026, as Barron’s reports. These expenditures have put pressure on Meta’s finances, with recent quarters showing liabilities and debt exceeding available cash, according to Barron’s.

The resolution of the Strike 3 lawsuit could have far-reaching effects on how AI training is conducted. If Meta wins, it could strengthen the position that companies are not responsible for unverified user actions on their networks, as discussed in

. On the other hand, a decision against Meta might require technology companies to implement stricter monitoring, raising costs and potentially hindering innovation.