Compiling Liability: The Copyright Dimensions of Algorithmic Social Media Timelines and Section 230 Implications

Generative AI certainly challenges our traditional notions of intellectual property rights, especially within copyright law. Novel AI tools such as Midjourney and Chat GPT raise new questions about whether works created with artificially intelligent machines can be copyrighted. While the U.S. Copyright Office announced that works created by AI without human intervention or involvement still cannot be copyrighted, there remains ample grey area. One such area concerns the output of algorithms. While algorithms are central to machine learning and artificial intelligence, their underlying equation is designed by a human. A human sets the algorithm’s parameters and is often engaged in an iterative process with the algorithm’s output, ensuring the result accords with the human’s underlying objective. This level of human involvement is notably absent when a machine creates an intricate piece of art in response to a one sentence prompt. 

If a causal nexus exists between an algorithm’s output and human authorship, then it stands to reason that this output can be copyrighted. And if this is the case, then we need to ask how might social media companies – arguably the algorithm’s most ubiquitous purveyor, be implicated by this conclusion? Algorithms are at the core of most social media companies’ business models. These proprietary equations are designed to take in massive amounts of user-generated inputs and, in turn, spit out highly tailored, original timelines. These timelines, also called feeds, are often comprised of various third-party content that aims to keep users engaged with the platform. Increased user engagement means more advertising revenue and, therefore, more profit. 

Unfortunately, despite numerous accusations that these algorithmically generated timelines are causing harm, including cases linking them to terrorist activity and increased rates of teen depression, social media companies remain legally immune from responsibility for alleged harms caused by their products. This immunity stems from the protections offered under Section 230 of the Communications Decency Act. Section 230(c)(1) specifies that service providers and users may not “be treated as the publisher or speaker of any information provided by another information content provider.” Therefore, Section 230 differentiates between content creators and those who facilitate access to the content, granting immunity to the latter. The key in determining the applicability of Section 230 immunity is whether the service provider played a role in developing content blamed for causing harm.

Now, you might be wondering how copyright law is relevant to Section 230 immunity. So far, courts have applied a neutral tool analysis to claims related to harms caused by algorithmically generated content, stating that using an algorithm to decide which information is published is akin to publishers using their editorial discretion to decide which articles to publish – activity protected under Section 230. However, one may argue that social media timelines are not just stand-alone snapshots of third-party content, but instead are copyrightable compilations that have been selected, coordinated, and arranged by each company. By treating algorithmically generated timelines as a compilation, it is easier to see how a timeline could be considered a developed work exempt from Section 230 immunity.

Compilations of preexisting works (also known as ‘collective works’) may . . . be copyrightable if the materials are selected, coordinated, or arranged in such a way that the resulting work as a whole constitutes a new work.

 U.S. Copyright Office

We need to look no further than the principles of copyright to reach this conclusion. According to the U.S. Copyright Office, “[C]ompilations of preexisting works (also known as ‘collective works’) may . . . be copyrightable if the materials are selected, coordinated, or arranged in such a way that the resulting work as a whole constitutes a new work.” Accordingly, “so long as [the compilations] are made independently by the compiler and entail a minimal degree of creativity . . . [c]ongress may protect such compilations through the copyright laws.”  Social media companies, using algorithms, are doing just this. Their equations are entirely original, so much so that they take great care to keep their algorithms secret from the rest of the world. And the timelines that are produced well exceed the low bar of minimal creativity needed for copyright protection. For example, if someone loves consuming Barbie related content, an algorithm will create a timeline significantly composed of such posts. However, if someone else is interested in woodworking, and engages with woodworking content more than they would with Barbie, then their timeline will be arranged accordingly.  If you print out both users’ timelines from the last 30 days and compare them side-by-side, it would be glaringly obvious that the social media company had a significant role in developing the two totally different works. Each timeline would be the result of an algorithmically curated process carefully selected, coordinated, and arranged in a way entirely unique to the social media company.

Why does this matter? Viewing social media timelines as compilations could have monumental effects on the current liability regime. Because compilations require a certain level of originality and creativity, if a timeline is deemed a compilation, then there is a strong argument that the timeline is a developed work under Section 230. If so, the author of the timeline – the social media company – would not be afforded the same immunity that a typical online publisher of third-party content enjoys. Thus, social media companies would no longer be able to hide behind Section 230 immunity. Instead, there would finally be a way to hold these companies liable for their products’ harm – something society desperately needs.

Casey Scofield

Casey Scofield is a 2L at the University of North Carolina School of Law.