Targeting Children: Liability for Algorithmic Recommendations

We live in the algorithmic society, characterized by massive digital surveillance and data collection by private companies exploiting human information vulnerabilities for profit. The infrastructure of free expression translates into an infrastructure of digital surveillance. This model, dubbed “surveillance capitalism,” includes massive personalized algorithmic targeting that departs from human speakers, allowing a level of influence never witnessed before in scale, scope, or depth.<br><br>Personalized targeting can garner many benefits, as it allows individuals to find content that interests them without needing to invest energy in seeking out that content. However, personalized targeting can also cause tremendous harm because it can alter an individual’s perceptual, emotional, and social judgment. It can also promote the illegal sale of firearms and drugs, increase pedophile networks, and amplify incitement to terror. Most critically for this Article, personalized algorithmic recommendations often target material to children that a platform’s official policy forbids. Because children are particularly vulnerable and lack full decision-making capacity, unregulated targeting can lead to self-harm and unfortunately, has even cost children’s lives. Should intermediaries bear any liability for targeting children? Or rather, should they be immune for autonomous targeting by artificial intelligence algorithms?<br><br>In the wake of the recent U.S. Supreme Court decisions in <i>Twitter, Inc. v. Taamneh and Gonzalez v. Google,&nbsp;</i>a change in policy regarding targeting is more urgent than ever, especially in the context of targeting susceptible children. Accordingly, this Article argues that with the shift from human information to a data network connection in the algorithmic society, algorithmic targeting should be more regulated. After all, platforms are not just middlemen; algorithmic targeting differs from merely hosting content, and it is not the same as human speech. In fact, the design of the platform and the algorithmic targeting mechanism itself develop content and transform the context of the information. Therefore, intermediaries should be subject to liability for negligent&nbsp;design.<br><br>This Article demonstrates how social media platforms, aiming to increase traffic and enhance profits, deliberately use algorithms to target susceptible children. This Article makes the case for imposing civil liability on intermediaries for negligent design, thereby incentivizing intermediaries to restrict algorithmic targeting directed towards children. Subsequently, this Article reviews normative free speech considerations in imposing liability for targeting children and overviews judicial decisions regarding liability by analyzing and criticizing case law interpreting § 230 immunity.<br><br>This Article outlines a proposed framework and policy reform for a duty of care for targeting children. Finally, it addresses possible objections to liability for algorithmic targeting, the problem of demonstrating causal connection between targeting and harm, and First Amendment objections to the proposed framework. The Article concludes that liability alone will not be enough to protect children, and more comprehensive reforms are required.