Meta faces $350 million lawsuit over adult film downloads

Close-up editorial collage with a large Meta logo centered over balanced scales, a pixelated red rectangle implying censored video, and glowing network cables curling around, bright cyan and deep blue contrasted with warm amber highlights, clean studio background, no people, high clarity, medium framing

Meta filed a motion to dismiss a lawsuit claiming the company illegally torrented adult films to train artificial intelligence models. Strike 3 Holdings alleged Meta downloaded its copyrighted content through corporate IP addresses and a network of hidden IPs. The lawsuit sought damages that could exceed $350 million.

Company Claims Personal Use by Individuals

Meta argues the downloads happened for personal use, not AI training. According to Ars Technica, the company said the activity amounted to about 22 downloads per year over seven years. Meta called this pattern inconsistent with the massive datasets needed for AI training.

The downloads started in 2018, four years before Meta began research on generative video models. Meta noted its terms prohibit generating adult content. This makes AI training with such material implausible, the company argued.

Meta said tens of thousands of employees, contractors, visitors, and third parties access its network daily. The company cannot confirm any specific Meta employee made the downloads. The filing said guests, contractors, or repair workers could have been responsible.

Strike 3 Cited Stealth Network Claims

Strike 3 alleged Meta used a stealth network of 2,500 hidden IP addresses to conceal some downloads. Meta called this claim nonsensical. The company asked why it would hide some downloads but use traceable corporate IPs for hundreds of others.

One claim involved a Meta contractor downloading content at his father’s house. Meta noted the contractor worked as an automation engineer with no apparent role in sourcing AI training data. The company said the downloads suggest personal consumption, not corporate activity.

Court to Review Dismissal Request

Meta called the lawsuit based on guesswork and innuendo. A company spokesperson told Ars the claims are bogus. Meta emphasized it takes deliberate steps to avoid training AI on adult material.

The company argued Strike 3 provided no evidence Meta trained models on the content. Monitoring every file downloaded on its global network would be extraordinarily complex and invasive, Meta said. Strike 3 has two weeks to respond to the motion.

Total
0
Shares
Previous Post
Close-up portrait of Jensen Huang in a black leather jacket, centered and brightly lit, with a luminous emerald Nvidia logo behind him, a glowing gold AI chip below, and faint neutral silhouettes of Donald Trump and Xi Jinping on opposite sides, vibrant warm gold against cool green and charcoal tones, studio-clean background with soft depth of field and medium-tight framing.

Nvidia becomes first company worth $5 trillion

Next Post
Editorial collage with two neutral, high-detail portraits of Sam Altman and Sarah Friar in the foreground, a large OpenAI swirl logo glowing behind them, golden upward arrow and stacked metallic coins forming a rising arc, bright warm golds against cool electric blues and deep purples, shallow depth of field, clean studio look, no text

OpenAI plans public stock sale that could reach $1 trillion

Related Posts