Meta Seeks Dismissal of Lawsuit Alleging Unauthorized Use of Adult Content for AI Training
Background of the Allegations
Meta is currently facing a lawsuit accusing the company of illicitly downloading adult films via torrent networks to train its artificial intelligence systems. The plaintiff, Strike 3 Holdings, claims to have identified unauthorized downloads of approximately 2,400 adult videos originating from Meta’s corporate IP addresses. Additionally, Strike 3 alleges that Meta attempted to obscure some of these downloads through a covert network of around 2,500 concealed IP addresses.
Meta’s Legal Response: Challenging the Claims
In response, Meta has filed a motion to dismiss the case, arguing that the accusations are based on speculation and lack concrete evidence. The company’s legal team describes Strike 3 as a “copyright troll” known for pursuing aggressive and extortionate litigation. Meta insists that there is no proof linking the corporation directly to the alleged downloads or demonstrating that any adult content was intentionally used to train its AI models.
Scope and Nature of the Downloads
The alleged torrenting activity reportedly spans a period of seven years, starting in 2018. Meta emphasizes that its internal policies explicitly prohibit the creation or use of adult content within its AI training datasets, which contradicts the premise that such material would be valuable or even permitted for AI development. Unlike other cases where authors’ works are included in massive datasets, Meta points out that the volume of downloads attributed to its IP addresses averages only about 22 per year-far too limited to constitute a systematic effort to train AI on this content.
Uncertainty Over the Source of Downloads
Meta highlights the difficulty in attributing the downloads to specific employees or contractors. The company notes that thousands of employees, visitors, contractors, and third-party vendors access its network daily, making it impossible to definitively identify who was responsible for the alleged activity. Meta also disclosed that one contractor downloaded adult content from his personal residence, but these actions were characterized as personal use unrelated to any AI training initiatives.
Questioning the “Stealth” Network Allegation
Meta finds the claim of a “stealth” network of hidden IP addresses particularly perplexing. The company questions why it would allegedly conceal some downloads while simultaneously using easily traceable corporate IPs for hundreds of others. This inconsistency, Meta argues, undermines the credibility of Strike 3’s assertions.
Challenges in Monitoring Network Activity
Addressing the issue of network oversight, Meta explains that monitoring every file downloaded across its vast global infrastructure is a complex and intrusive task. The company maintains that it employs reasonable measures to detect unauthorized activity but cannot be held liable for every individual’s actions on its network. Meta’s legal team references precedent indicating that only basic monitoring obligations are required.
Implications for AI Content Policies and Future Litigation
Beyond contesting potential damages, Meta’s defense underscores its commitment to preventing the use of explicit content in AI training. The company stresses that it actively implements safeguards to avoid incorporating adult material into its AI models, aligning with increasing regulatory scrutiny over AI-generated content. Meta’s spokesperson reiterated that no evidence has been presented to show that their AI systems were trained on Strike 3’s copyrighted adult content.
Next Steps in the Legal Battle
Strike 3 has a limited window of two weeks to respond to Meta’s motion to dismiss. The outcome of this case could have significant ramifications for how tech companies manage copyrighted content in AI training datasets, especially as the industry faces growing calls for transparency and ethical standards.
