And now, Eriq analyzes the new Texas lawsuit targeting Netflix algorithms…

Last month, Ken Paxton, the Texas attorney general now running for Senate with Trump’s enthusiastic backing, filed a lawsuit against Netflix with an ominous opening—“When you watch Netflix, Netflix watches you.” His thesis was even more sinister. According to Paxton, Netflix deliberately designed its platform to keep children glued to their screens. “Netflix uses dark patterns to do this discreetly,” the complaint alleges. “Dark patterns are subtle features engineered to manipulate users to take the actions Netflix wants them to take.” Unaware of these subterfuges, Paxton writes, “Texans signed up in droves and handed their children the remote.”

A MESSAGE FROM OUR SPONSOR

For your awards consideration, HBO Max presents INDUSTRY. In Season 4, Harper and Yasmin are on opposing sides of a high-stakes financial deal that could bring them both down. Who will finish on top?

Sure, but… what entertainment company doesn’t try to hook viewers? For decades, that sort of thing was considered squarely within the realm of First Amendment protection. Hollywood’s business model has always been to persuade people to watch its stuff. And when states have wandered too far into trying to regulate these tactics, courts have generally swatted them away. Remember when former California Gov. Arnold Schwarzenegger tried to restrict the sale of violent video games to minors? The Supreme Court ultimately held that law unconstitutional, reaffirming that entertainment products do not lose First Amendment protection just because kids like them.

So what makes this different? Netflix, for its part, argues that Paxton’s lawsuit rests on a misunderstanding of how its platform actually works. On Monday, the company made a special appearance in a Collin County courtroom to argue that Texas lacks jurisdiction because the policies under attack were devised and overseen by executives in California. As for the broader allegations, Netflix has so far kept its powder dry. But it told me the lawsuit “lacks merit and is based on inaccurate and distorted information.” The company also noted that autoplay can be disabled, and the platform allows parents to retain control over children’s viewing.

Regardless, Paxton’s strategy is easy to spot. He is plainly trying to exploit what many regulators and plaintiffs’ lawyers now see as an opening: product-design liability. After years of hitting Section 230 and First Amendment roadblocks, the argument is no longer that the content itself is harmful. Now, it’s that platforms have engineered systems to exploit vulnerabilities, maximize engagement, and encourage compulsive use. The legal target is shifting from the actual show, song, video, or post to the machinery controlling how it is served up.

The distinction is gaining traction in courtrooms. Earlier this year, a Los Angeles jury found Meta and YouTube negligent, awarding $6 million to a 20-year-old California woman, represented by Mark Lanier, who claimed that the two platforms were designed to hook young users, creating an “engineered addiction” that harmed mental health. That followed a $375 million New Mexico verdict against Meta over claims that the company misled users about child safety and failed to protect young people from exploitation. (The state had sought $2 billion.) It was only a matter of time before the theory migrated to streaming services.

If Paxton’s lawsuit represents one front in the emerging war over content in the digital age, another battle is already underway. This Wednesday, a federal courthouse in San Jose will host a hearing that could prove even more consequential. Before regulators can rule on how Netflix is permitted to engage viewers, they first have to answer a more fundamental question: What exactly counts as editorial judgment in the algorithmic era?

At issue is California’s SB 976, the “Protecting Our Kids from Social Media Addiction Act,” scheduled to take effect next year, which prohibits platforms from serving personalized feeds to minors absent parental consent. Rather than wait, TikTok is asking Judge Edward Davila to block enforcement of the law now. What makes this hearing particularly fascinating is that the court has zeroed in on asking the parties to brief him on how much human involvement in an algorithm is required to qualify a personalized feed as protected speech.

The inquiry has already staked out some new legal territory. For starters, TikTok is effectively opening the black box of its recommendation engine, submitting declarations from product and design executives who describe the human judgments, policies, and editorial choices that influence what users ultimately see. That algorithm took on near-mythical status as the app’s adoption exploded in the past half decade—a reputation the company hasn’t exactly discouraged. A few years ago, when lawmakers were consumed with fears of covert Chinese influence over American teenagers, it would have been hard to imagine TikTok voluntarily inviting a court to look under the hood. Now TikTok product manager Joey Zhou is proclaiming, “The sophisticated predictive models of the recommendation do not remotely operate through random chance, like the roll of a die. Quite the opposite.”

A MESSAGE FROM OUR SPONSOR

For your awards consideration, HBO Max presents THE CHAIR COMPANY. After an embarrassing incident at work, a man (Emmy® winner Tim Robinson) finds himself investigating a far-reaching conspiracy. Don't miss the series The Daily Beast called a “MASTERPIECE”. THE CHAIR COMPANY is now streaming on HBO MAX.

The implications extend well beyond TikTok. The question goes to the heart of First Amendment doctrine as algorithms increasingly determine not only how content is delivered, but how it’s created. At what point does human involvement become so attenuated that constitutional protection begins to fade? That is a question that particularly interested Justice Amy Coney Barrett during arguments in Moody v. NetChoice, the Supreme Court’s 2024 attempt to grapple with Florida and Texas laws aimed at preventing platforms from allegedly “censoring” conservative viewpoints. Barrett’s concern was not academic: If a platform’s ranking system becomes less a tool for executing human editorial policy and more an autonomous prediction engine, courts will have to decide whether the First Amendment follows the human intention, the machine output, or neither.

The result in Moody was something of a muddle. The Court broadly reaffirmed that social media companies possess First Amendment rights to organize, prioritize, and exclude content. But it offered little guidance on the harder questions involving algorithmic recommendations, engagement optimization, and product design. Those issues were largely kicked down the road.

That road now leads straight to Davila’s courtroom. And unlike the culture-war framing that dominated Moody, this case centers on the fundamental question of whose speech is it, exactly, when what appears on a screen is the product of human judgment, machine learning, and automated optimization working in tandem.

California Attorney General Rob Bonta, defending SB 976, points to scale. TikTok’s algorithms generate billions of unique feeds. Even conceding that the company makes broad policy decisions about what content gets promoted or suppressed, the state argues that TikTok is largely ignorant of the specific output appearing on any individual screen—and at that level, the recommendation engine is no longer merely implementing human editorial choices. The algorithms, California argues, are “making ‘decisions’ of their own, based on billions of inputs and considerations that are far beyond the comprehension of any human being.”

TikTok, represented by the veteran Hollywood litigator Daniel Petrocelli, sees the matter differently. He writes in a brief: “There is no basis in the First Amendment for distinguishing between human control over how speech is compiled and human execution of compilation decisions at the individual level. The Constitution, TikTok argues, protects editorial judgment whether exercised by a newspaper editor arranging a front page or engineers designing a system that arranges content at scale. The company also presses the slippery-slope concern: If First Amendment protection requires a human personally reviewing each piece of delivered content, personalized feeds on Apple News and personalized playlists on Spotify become equally vulnerable to regulation.

There is obviously an awkward aspect to TikTok’s position. To win First Amendment protection, the company must emphasize the policies, teams, engineers, and product judgments shaping what users see. That helps in Davila’s courtroom. But it is not always how platforms prefer to describe themselves when facing product-liability claims, consumer-protection suits, or especially political accusations that they manipulate public discourse. The more human the feed becomes for constitutional purposes, the less plausible it may be to describe the platform elsewhere as merely reflecting user choice. In other words, the defense strategy may win today’s fight while making others harder to win. But that’s a story for another day.

Thanks, Eriq. See everyone tomorrow.

Matt

Need help? Review our FAQ page or contact us for assistance. For brand partnerships, email ads@puck.news.

You received this email because you signed up to receive emails from Puck, or as part of your Puck account associated with {{customer.email}}. To stop receiving this newsletter and/or manage all your email preferences, click here.

Puck is published by Heat Media LLC. 107 Greenwich St., New York, NY 10006