This is an expanded version of a post that first appeared on Scott Graham's Skilled in the Art IP briefing.

We had the Rumble in the Copyright Jungle this morning in U.S. Magistrate Judge Alex Tse's Northern District courtroom. Stanford Law School professor Mark Lemley and Orrick, Herrington & Sutcliffe partner Annette Hurst debated whether Amazon.com's software algorithms are evolved to the point that they exercise the human-like volition necessary for copyright liability when they choose photographs to appear on product pages.

Hurst, who represents Williams-Somona Inc., warned that a ruling for Amazon will put the law on a dystopian path where no one can be held responsible for killings by automated vehicles and AI-piloted drones. Lemley, representing Amazon, rejected the premise. "This is not a rogue AI case. We are not in that world," the Durie Tangri partner said. "I can imagine we might end up in such a world at some point, but we are not there now."

Williams-Sonoma sued Amazon 18 months ago, alleging the tech giant misleads consumers into believing that third-party resellers of Williams-Sonoma merchandise are authorized dealers. The company also accuses Amazon of infringing several design patents. Williams-Sonoma then added copyright claims this spring, alleging that the resellers are uploading to Amazon photos that are professionally shot for and copyrighted by Williams-Sonoma.

Normally, an internet service provider is shielded from liability by the Digital Millennium Copyright Act when it merely hosts a third-party's infringing content. Williams-Sonoma contends that Amazon has developed sophisticated AI-based software known among employees as "Frankenstein" that chooses from among the photos. "The reason they call them Frankenstein, your honor, is because Amazon is Dr. Frankenstein creating these monsters," Hurst said. Frankenstein is "picking winners, and it's using others' copyrighted content to do so, and directly financially benefiting by taking a cut."

Amazon is asking Tse to dismiss the copyright claims. It says the company is doing nothing out of the ordinary and the DMCA safe harbor should apply. Websites like YouTube, Facebook and Instagram use software that processes, ranks and prioritizes the content that people post to it, said Lemley. "Were Williams-Sonoma's proposal to make new law adopted, the internet as we know it would be impossible," he said, "because no one could host content posted by third parties without fear of crushing liability."

Tse said he's going to be guided by Justice Antonin Scalia's dissent in ABC v. Aereo, where he wrote that a service provider who is "totally indifferent to the material" is shielded from liability. Tse added that it's "hard for me to disregard the allegations in the complaint" at this early stage of the proceedings.

Lemley said courts have consistently held machines less liable than human beings, "because our effort to find a volitional actor is an effort to find a human being at the end of the day. There is nothing in this complaint that alleges any human intervention, any human decision-making in this process."

Hurst called that a mischaracterization of her argument. "The important distinction here is not between a computer and a human, it's between automatic and manual. In the past we have understood manual to be human, because we haven't had software agents that were capable of exercising the kind of specific discretion" to be considered manual.

Software has now evolved beyond that point, and the law should evolve with it, Hurst said. "If the rule is going to be they're absolved of liability simply because they're a computer, then we will have automated vehicles killing people, and no ability to hold anyone responsible. We'll have drones killing people and no ability to hold anyone responsible.

"The entity, Amazon, who put that software into the stream of commerce is responsible for its act," she argued.

Lemley said that is far from this case. "There is absolutely nothing in the complaint" that suggests AI run amok, he said. "What the complaint suggests is that when two photos are uploaded for the same product, the algorithm makes a choice as to which photo will be displayed," he said. "It says nothing about individualized, manual, human-like decision making."

Tse did not tip his hand on the bench, instead promising to issue a written order.