Blocking as Regulating? Blacklisting Generative AI, 73 Am. U. L. Rev. (forthcoming 2024)
51 Pages Posted: 22 Aug 2022 Last revised: 17 Oct 2023
Date Written: August 16, 2022
Abstract
Policy-makers and administrations often play a critical role in regulating the use and advancement of technology. Their decisions can have significant positive impacts on the development, adoption, and use of various technologies, but also negative ones if they decide to limit or even block a certain technology. Italy’s temporary banning of the Generative Artificial Intelligence (GenAI) platform ChatGPT, much like New York City's Department of Education, continues recent years’ reliance of administrations and policy-makers on blacklisting and banning as significant tools in their arsenal. Yet, the concept of blocking via blacklisting or banning is not new. Historically, being a blacklisted entity meant having a negative reputation of an entity that cannot be trusted or has done something wrong. Regulatory blacklisting, based on a similar notion, enables lawmakers and public officials to enforce regulatory standards on businesses, individuals, organizations and even jurisdictions by using signaling economy and sanctions. Similarly, banning, which can also help regulators control unwanted technology or its popularity, has historically been used to prevent the use of undesired products or services by the public.
Blacklisting and banning are two similar terms that are often used interchangeably. The confusion between the terms happens mainly because typically, when persons or entities are blacklisted they are banned from using, benefiting, or profiting from something. Nevertheless, in the context of regulation, the terms are different and used for different purposes, and it is key to understand why, how, and what consequences are associated with each term, particularly when used in connection with innovative and disruptive technologies such as GenAI.
This Article explores using blacklisting as a regulatory tool while differentiating it from the practice of banning. Doing so, this Article examines the impact of any related explicit and implicit sanctions, especially as such sanctions include reputational damage, increased hardship in getting credit, and the potential costs of “doing business.” Additionally, the Article makes the important distinction between getting blacklisted for participating in or advancing an illegal activity or failing to comply with binding legal requirements, as, for example, was the case of the crypto mixing service Tornado Cash, and getting blacklisted for moral or ethical reasons, as, for example, President Trump’s administration perceived the case with TikTok – the Chinese social platform – to be. It explains why distinguishing between the two matters for regulatory purposes in general, and for GenAI’s case in particular, and describes why sanctions do not always work. Finally, the Article makes policy recommendations that would help increase the effectiveness of blacklisting as a regulatory tool, while acknowledging the problematic nature of regulation by enforcement.
Keywords: Blacklisting, Law and Economics, Sanctions, Cryptocurencies, Regulation by Enforcement
JEL Classification: K00, K22, K42, M00
Suggested Citation: Suggested Citation