“They will only get in trouble if they summarize the story incorrectly and make it defamatory when it wasn’t before. That’s something they’re really going to be at risk for legally, especially if they don’t clearly state the source and people can’t easily access that source to check,” he said. “If Perplexity’s edits were what made the story defamatory, then 230 would not have included that, according to a series of case laws explaining it.”
In one case WIRED observed, Perplexity’s chatbot falsely claimed, despite a clear link to the original source, that WIRED had reported that a specific police officer in California had committed a crime. (“We have been upfront in saying that the answers will not be 100% accurate in all cases and may cause hallucinations,” Srinivas said in response to questions for a story we published earlier this week, “but a core aspect of our mission is continued improvement in accuracy and user experience.”)
“If you want to be formal, I think this is a set of statements that would pass muster with a proposal to refute a bunch of theories,” Grimmelmann said. Not saying they will win in the end, but if the facts prove what Forbes and WIRED, the police—a series of possible plaintiffs—say, then those are the things that if proven and the facts otherwise would be harmful to Perplexity, possibly giving rise to liability.”
Not all experts agree with Grimmelmann. Pam Samuelson, a professor of law and communication at UC Berkeley, wrote in an email that copyright infringement is “the use of another person’s expression in a way that impairs the author’s ability to receive compensation.” suitable for the value of unauthorized use. The original sentence is probably not a violation.”
Bhamati Viswanathan, a lecturer at New England Law, said that she suspects the brief passes the substantial similarity threshold typically needed for a successful infringement claim, although she doesn’t think the matter is over yet. . “It certainly wouldn’t pass the sniff test,” she wrote in an email. “I think just getting your case past the threshold for a motion to dismiss is enough — especially when you have all the indications that the actual content is being copied.”
In general, however, she argues that focusing on the narrow technical value of such claims may not be the right way to think about things, since tech companies can manipulate adapt their operations to respect the letter of outdated copyright law while still egregiously violating their intent. She believes that an entirely new regulatory framework may be needed to correct market distortions and advance the fundamental goals of US intellectual property law, among them allowing people to benefit financially from original creative work such as journalism so that they have an incentive to produce it. – theoretically, benefit society.
“In my view, there are strong arguments in favor of the intuition that generative AI relies on large-scale piracy,” she wrote. “The opening question is, where do we go from there? And the bigger question in the long term is how do we ensure that creators and the creative economy survive? Ironically, AI is teaching us that creativity is more valuable and in demand than ever. But even as we recognize this, we still see the potential to weaken and ultimately destroy the ecosystems that allow creators to make a living from their work. That is the conundrum we need to solve—not eventually but now.”
