AI and Online Child-Safety Dominate Agenda at San Francisco Digital Conference

In a one-year period in which over 20 lawsuits were filed alleging that AI chatbots were the cause of psychological harm and teen suicides; juries in Los Angeles and New Mexico awarded substantial damages against major tech companies; and teen social media bans were implemented in Australia and other countries around the world, the agenda at MLRC’s annual Legal Frontiers in Digital Media conference practically wrote itself. Our May 14^th conference brought together approximately 155 lawyers – some from across the globe – to San Francisco to share their perspectives on the shifting legal landscape that is being driven by concerns surrounding AI and child safety.

On a panel that extensively covered the K.G.M. trial against Meta and YouTube, in which a teenage social media user alleged that addictive features of the social media platforms negatively impacted her childhood development, we were very fortunate to have a New York Times reporter who was in the courtroom for the trial, Eli Tan. He explained how the plaintiff’s lawyer presented the case without reference to screenshots showing the content the plaintiff had been exposed to (to avoid Section 230 immunity) but rather by comparing social media features – like infinite scroll and push notifications – to the addictive properties of slot machines and cigarettes, using the platforms’ own internal documents to make the case that the companies knew their products were addictive. Based on Tan’s interviews with some of the jurors, it was the way the employees of the tech companies discussed their products in private that drove their finding of liability against the social media giants. Several members of the panel downplayed the significance of the verdict, including noted internet law scholar, Professor Eric Goldman, who pointed to serious flaws in pre-trial rulings with respect to the First Amendment and Section 230 and identified a series of grounds for appeal.

A separate panel was organized to discuss global child safety regulation, with a particular focus on Australia’s ban on teens under 16 from using social media platforms. We were extremely fortunate that panel organizer and moderator Janneke Slöetjes was able to persuade Sydney, Australia-based lawyer, Tanvi Mehta Krensel, to make the 13+ hour flight to join us in San Francisco. Krensel, along with the other panelists, Reddit’s Sunni Yuen and Roblox’s Noah Bialos, agreed that the push toward social media bans, which started in Australia and spread throughout the world, gained momentum through the popularity of the book, The Anxious Generation, by Jonathan Haidt. The book argues that the introduction of smartphones and social media apps to teenagers has led to myriad social and psychological harms to teens. In Australia, social media companies are required to take “reasonable steps” to block teens under 16, but the panel noted how differences in the laws and enforcement of such age-gating requirements around the globe, with no clear consensus emerging, has made it increasingly challenging for the tech firms to comply with regulation in each jurisdiction.

At the crossroads of concerns over child safety and AI, another panel discussed the wave of recent product liability claims against companies that have marketed AI chatbots, some in cases where it has been alleged that chatbots contributed to suicides (as well as other psychological harms to both minors and adults). According to Mashable reporter, Rebecca Ruiz, a common pattern has been that teen communications with chatbots begin as requests for help with homework, but the conversations delve off into other areas, sometimes with tragic results, such as when the chatbot offers help with plans for self-harm or harm to others. Given these serious concerns, an interesting question addressed by the panel is to what extent such communications are still protected speech under the First Amendment. Professor Eugene Volokh argued that chatbot communications should largely receive the same constitutional protection as controversial music, such as Ozzy Osbourne’s “Suicide Solution,” and violent video games, which some have alleged to have led to tragic harm. Another scholar, Mackenzie Austin, argued that AI outputs are not protected speech because the speaker, in the case of an AI chatbot, doesn’t understand what it is expressing at the moment it is being communicated.

Other sessions addressed advising clients on developing AI products (in a panel skillfully led by Fenwick’s Kimberly Culp), the implications of the Supreme Court’s recent copyright decision in Cox v. Sony Music (expertly presented by UC Berkeley Law Professor Pamela Samuelson) and EU regulatory and liability developments in a fireside chat with conference regulars Remy Chavannes and Daphne Keller.

As regular attendees of our digital conference know, the conversations among colleagues that occur during the breaks are just as important as the content. To facilitate attendees “breaking bread” together, we were pleased to once again have Microsoft sponsor lunch and Google sponsor our evening reception. Other sponsors that helped make our conference possible were Alger Resolutions, Anthropic, Arnold & Porter, Davis Wright Tremaine, Fenwick, FIRE, Jassy Vick Carolan, Kilpatrick, Meta, Munger Tolles & Olson, Pillsbury, and Williams & Connolly. The conference will once again be held next year, at the Mission Bay Conference Center, on Thursday, May 13, 2027.