Meta, the parent company of Facebook and Instagram, is facing a $2.8 billion lawsuit after being accused of illegally downloading and distributing thousands of adult films to train its artificial intelligence (AI) models. Strike 3 Holdings, a major adult content producer, claims that Meta exploited the BitTorrent protocol to rapidly obtain and share high-resolution adult videos, sometimes even on the day of release. This alleged activity not only violates copyright but also gives Meta an unfair advantage in accumulating diverse training data for AI development.

According to the lawsuit, Meta intentionally distributed these films for extended periods, leveraging BitTorrent’s “tit-for-tat” sharing mechanism to access a wide range of content more efficiently. The suit contends this approach allowed Meta to gather huge quantities of creative material while bypassing traditional licensing channels and intellectual property protections. Moreover, the complaint suggests these tactics could have endangered minors, as distributing adult content without proper age verification is especially problematic in regions with strict regulations.

This case has reignited debates surrounding the use of copyrighted and sensitive content in AI training. While courts have sometimes ruled that using large text databases for AI training is “transformative” and protected under certain conditions, the line becomes much blurrier when the source material is unique, high-value, and not intended for public distribution. The outcome of this lawsuit could set significant precedents on how tech giants can obtain and use creative media to fuel the next generation of AI systems.

These events highlight the risks faced by content creators in the digital age. As AI models become ever more advanced, the tension between the demand for high-quality training data and the rights of original creators will likely intensify, making intellectual property and digital publishing strategies more crucial than ever.