Looks like Meta is hoping the recent Supreme Court ruling that found Internet service providers aren’t liable for piracy on their networks will help the social media giant dodge liability claims over its torrenting of AI training data.
Last week, Meta filed a statement in a lawsuit that alleged that Meta should be liable under copyright law for contributory infringement simply because the company knows how torrenting works. By seeding perhaps 80 terabytes of pirated works, the company allegedly knew it was inducing infringement by allowing uploads to help speed up its downloads, the plaintiffs, Entrepreneur Media, argued.
This contributory infringement claim is much easier to prove than a separate claim raised in a class action filed by book authors in Kadrey v. Meta, which alleged that Meta’s torrenting meant it was liable for a “distribution” claim of direct copyright infringement. TorrentFreak noted that the authors’ claim required evidence that Meta torrented an entire work, whereas the contributory infringement claim only depends on proving that Meta facilitated torrent transfers.
Since torrenting depends on swarms of users sharing fragments of files, authors in the class action were expected to struggle to show any of the disputed works were seeded in full. That’s why, for Meta, the contributory infringement claim’s survival could be damning, as a judge recently ruled that it can be added to the class action case, too.
Meta’s strategy to avoid losing both cases now appears to hinge at least partly on courts agreeing that the Supreme Court’s ruling in the Cox case drew a bright red line in a way that benefits Meta when clarifying the standards for contributory infringement. In its statement, Meta said it would soon file a supplemental brief explaining why the ruling would support its motion to dismiss the Entrepreneur Media case.
Meta’s new tactic to defeat torrenting claims
It’s unclear exactly what Meta will argue at this point, but Ars previously explained that the Supreme Court found that prior court precedent only supports two theories of secondary liability when it comes to contributory infringement.
One theory, which applied to ISPs like Cox, finds a company “not liable as a copyright infringer for merely providing a service to the general public with knowledge that it will be used by some to infringe copyrights.” On its face, this does not seem to describe Meta’s torrenting, which relied on BitTorrent to provide the service that was distributing Meta’s seeds, rather than Meta.
The other theory, which Meta seems likely to invoke, finds a company not liable unless plaintiffs can prove they “affirmatively induced the infringement.”
In Meta’s motion to dismiss the Entrepreneur Media case, the company has already argued that plaintiffs’ complaint “does not allege any facts identifying whom Meta allegedly shared” its training data with. It also doesn’t allege that “Meta had actual knowledge of specific infringing acts by third parties” or that “Meta could have taken simple measures to prevent such acts,” Meta argued. Therefore, Meta could not have induced infringement, the company has argued.
However, plaintiffs in both cases have insisted that Meta knows how torrenting works, and that bare assertion should be enough. Opposing the motion to dismiss in the Entrepreneur Media case, the magazine’s legal team argued that “third-party infringement is inherent” to torrenting, and “any user of BitTorrent understands that downloading requires uploading—that is the system’s core design.”
“Meta’s knowledge was not ‘generalized’ awareness of potential infringement; it was certain knowledge of an unavoidable consequence,” Entrepreneur Media argued.
But Meta is not the only one signaling that the SCOTUS ruling will favor AI firms like Meta who are accused of torrenting. Plaintiffs in both cases opposing Meta’s torrenting may have viewed the ruling as bad news after reading Justice Sonia Sotomayor’s concurrence.
Sotomayor ultimately agreed with the majority but warned that the “majority’s limiting of secondary liability here dismantles the statutory incentive structure that Congress created” in the DMCA. She suggested the ruling may ignore “other common-law theories of such liability, like aiding and abetting,” that “could apply in the copyright context.”
On his blog, Internet law professor Eric Goldman noted that the standard that the majority upheld “should favor future defendants,” while copyright owners will likely “invest a lot of money to try to water it down and undermine it.”
Meta did not immediately respond to Ars’ request to comment.
Judge accuses lawyers of bashing Meta
Although Meta clearly took comfort in the Supreme Court ruling, the company was likely disappointed that the contributory infringement claim was added to the class action case. The company may have been banking on the request coming too late for the judge to approve it.
In an order granting authors’ request to add the claim, US District Judge Vince Chhabria agreed that authors’ lawyers did not necessarily deserve the win in the fight. He scolded authors’ lawyers who—“particularly since the Boies Schiller firm came in and took the lead”—have seemed more intent on “bashing” Meta than shoring up their legal arguments.
It was “vexing,” Chhabria wrote, that authors could have added this claim as early as November 2024. Their lawyers now offer a “lame excuse” for the late update, he said, while seeking to amend their complaint a fifth time, which ordinarily would have been unacceptable so late into litigation that started in 2023.
Lawyers offered “no excuse for neglecting to add a contributory infringement claim” sooner, Chhabria said. Instead, they offered a “bunch of doubletalk” about Meta supposedly delaying discovery.
It irked the judge that they claimed that Meta had “newly produced evidence” that “provides a level of detail that was previously missing in this case,” Chhabria wrote. If that were “really true,” the judge suggested that lawyers wouldn’t have had enough discovery to add their “distribution” claim in 2024, either.
He accused the lawyers of covering up their own failure to flesh out their arguments earlier on in the case.
“Rather than blaming Meta for producing discovery late, the plaintiffs’ lawyers should have been candid with the Court, explaining that they missed an issue in a case of first impression and asking for permission to correct it because doing so would serve the interests of the proposed class without interfering with the case schedule or otherwise unfairly prejudicing Meta,” Chhabria wrote.
Later, the judge said this was “part of an ongoing pattern,” where authors’ lawyers try to “distract” from their own “mistakes” by “bashing Meta.”
“They seem so intent on bashing Meta that they are unable to exercise proper judgment about how to represent the interests of their clients and the proposed class members,” Chhabria wrote.
A lead lawyer from the Boies Schiller firm did not immediately respond to Ars’ request to comment.
Meta request changes luck for authors
The request from authors’ legal team was so delayed that it should have been “an easy deny,” Chhabria said. However, lawyers “lucked” into a win this time, mostly because Meta had requested that the Entrepreneur Media case align its schedule with the class action. That linked discovery between the two cases.
Because Meta was going to face the contributory infringement claim anyway, it would not prejudice Meta to require discovery on the issue so late in the class action, Chhabria wrote.
Further, “denying the motion to add the contributory infringement claim could potentially harm the interests of the proposed class members,” Chhabria said. If the class action proceeded without the claim, members would be barred from ever raising it, even if the Entrepreneur Media case ruling went against Meta, the judge noted.
“There is a serious concern that the interests of the absent class members would be harmed, through no fault of their own,” Chhabria wrote, while noting that he granted authors’ request “reluctantly.”
“On the flip side,” Chhabria said, adding the claim to the class action basically “meant that if the named plaintiffs obtained summary judgment and subsequently obtained class certification, proposed class members would know, when deciding whether to opt out of the class, that they had essentially already won.”
Chhabria’s ruling perhaps incentivizes Meta to dodge these claims as fast as possible. He noted that Meta faces no discovery in the class action “until plaintiffs can get past summary judgment on the distribution and contributory infringement claims.”
Moving forward, authors may feel somewhat more optimistic that they could get a partial win. Chhabria explained the standard for contributory infringement as a lower bar, proving that Meta was “facilitating copyright infringement by third parties by uploading protected works onto the torrenting network.”
Yet authors can’t be sure, since looming on the horizon, Meta is drafting a filing based on the Supreme Court ruling that could change the game.
Already Meta is seemingly willing to make any argument to escape consequences for torrenting. It’s continuing to argue that the number of works at dispute is a small fraction of the total data that was torrented. And it has even claimed that there’s no way to prove that Meta ever knew that torrenting required uploading.
However, if Meta loses at the summary judgment stage, authors are ready to argue that none of Meta’s internal discussions of torrenting should be privileged. If that discovery request is eventually granted, it could finally expose who exactly at Meta approved the torrenting and how well did they understand how BitTorrent works.
