The past few years have seen increasing calls for regulation of large social media platforms, and several states have recently enacted laws regulating their content moderation, promotion, and recommendation practices. But if those platforms are exercising editorial discretion when carrying out these tasks, many of the regulations will run into constitutional concerns: the First Amendment protects the “exercise of editorial control and judgment” by publishers over their choice of content and how it is presented. However, the editorial operation of social media platforms differs significantly from traditional media, most importantly in the use of artificial intelligence (AI) for editorial decision-making. While courts have thus far not given much attention to the implications of this use, this Note argues that using AI for editorial decision-making should not be entitled to the same protection as a human decision-maker. After introducing how social media platforms employ AI in their editorial operations, it outlines the foundational values of democratic self-governance, the marketplace of ideas, and autonomy underlying the First Amendment, and assesses how the use of AI impacts those values. The Note concludes that the ability of AI to manipulate human behavior and preferences, combined with the delegation of decisional autonomy from humans to AI, harms the foundational First Amendment values. Therefore, the use of AI is not deserving of the same protection as human editorial decision-making.
Machine Manipulation: Why an AI Editor Does Not Serve First Amendment Values,
U. Colo. L. Rev.
Available at: https://scholar.law.colorado.edu/lawreview/vol95/iss1/7