In Latin, the word ‘apocalypse’ means ‘to unveil.’ That is to uncover, to disclose, to reveal knowledge. With the advent of the internet, information that would have previously required consultation with an expert is now accessible to all. We are amid an ‘Apocalypse of Knowledge,’ called the information age, facilitated by search engines and now by artificial intelligence (AI). As the lyric goes, “[t]he people bowed and prayed to the neon god they made.”1 In other words, pray to God or, with head bowed down to a cellphone screen, type anything and receive a response from the all-knowing LSTM neural net language model 4.0.2 

Regardless of the spiritual implications of this technology, a new era has certainly dawned. The veil has been lifted, the truth is seen, and Pandora’s box cannot be closed once it is open. In this new age of enlightenment, where nearly everyone has access to highly efficient informational and intelligence resources, what sort of legal protections must be afforded to normal human beings to secure their access to AI technology in a world also populated by immortal corporations, kleptocracies, and total informational awareness? 

Thankfully, usage of language learning model AIs, like ChatGPT, are likely Constitutionally protected as free speech and will be difficult for the government to override under strict scrutiny.3 More generally, the government will have a challenging time arguing against democratic access to advanced AI technologies for the following reasons: (1) Advanced AI, like ChatGPT, is educationally useful, so a lack of access thereto could result in educational inequity.4,5, (2) Disparate access willalso afford only certain groups the utility of AI for innovation and research.6 (3) Democratic access to AI will help maximize civic and democratic participation as far as the technology can improve expressional capacity.7 (4) Democratic access to AI will likewise prevent bias in the marketplace of ideas resultant from sophistry and the expression of linguistically superior opinions generated by those with unfair access to the technology.8,9, (5) Democratic access will drive economic growth and job creation because of resultant new industries and boosted productivity.10 (6) AI further allows for the counteracting of artificial narratives generated by bot farms, as well as nefarious people or organizations, since AI can help users navigate complex arguments, identify logical fallacies, and compose effective counterarguments.11 (7) AI also allows for greater access to expert-level advice in all domains of human knowledge and even medical and financial guidance, especially for those of lower socioeconomic status.12,13, (8) Furthermore, just like how access to the internet has democratized access to information, democratizing access to AI will empower the public to make better informed decisions, gain knowledge and challenge established narratives.14 (9) In that light, democratic access to AI will also generally level the playing field by giving ordinary people access to the same informational and intelligence resources that have been historically reserved for governments and large corporations, thereby reinforcing individual liberties against already ingrained, or emerging, hegemonic powers and ideologies.15

From a free speech perspective, prohibiting access to AI would be harmful to democracy for two primary reasons: First, it would unconstitutionally limit expression by people that do not have access to AI technology and, second, it would restrict others from hearing speech that would have otherwise been communicated, i.e., the right to listen.16 Consequently, a prohibition against AI technologies, like language learning models, would unconstitutionally hamper social interest in the free marketplace of ideas.17 The great experiment of American liberty is reliant upon the free expression of opposing, contradictory, and imperfect viewpoints.18 Essentially, as AI enters the marketplace, the promotion of information generated by AI will suppress information not generated by it due to the technology’s innate amplifying power.19,20,21,, Therefore, equitable access must be afforded to all so that the marketplace is not thereby undermined. 

In comparison, imagine that in an auditorium full of shouting citizens, only a few people have access to bullhorns; only those with access would have their speech amplified. The Supreme Court found that such wanton abuse of amplification devices may be prohibited to preserve public order or to prevent unreasonable interference.22 “Those who desire to broadcast can hardly acquiesce in a requirement to modulate their sounds to a pitch that would not rise above other street noises, nor would they deem a restriction to sparsely used localities or to hours after work and before sleep — say 6 to 9 p. m. — sufficient for the exercise of their claimed privilege.”23 The point being that those who have access to AI technologies will inherently use them to the detriment of others, because the technology is generally superior compared to equivalent organic means.24 However, in the case of bullhorns, this inherent inequality can be easily remedied by a city ordinance, but when the technology of amplification is entirely non-physical and nebulously contained on the internet, prohibition becomes extremely difficult. Therefore, universal access to AI technologies must be preserved and any attempt at prohibition must be struck down. 

References:

1 SIMON, P., & GARFUNKEL, A., The Sound of Silence, on WEDNESDAY MORNING, 3 A.M. (Columbia 1963).

2 Alexandru Hutanu, How ChatGPT Works and AI, ML & NLP Fundamentals, Pentalog, (Feb. 8, 2023), https://www.pentalog.com/blog/tech-trends/chatgpt-fundamentals/

3 Texas v. Johnson, 491 U.S. 397 (1989)

4 Yongjun Xu et al., Artificial Intelligence: A Powerful Paradigm for Scientific Research, 2 The Innovation 100179 (2021), https://doi.org/10.1016/j.xinn.2021.100179

5 U.S. Dep’t of Educ., Office of Educ. Tech., Artificial Intelligence and the Future of Teaching and Learning: Insights and Recommendations (May 2023), https://www2.ed.gov/documents/ai-report/ai-report.pdf

6 Ibid.

7 Hélène Landemore, Fostering a More Inclusive Democracy with AI, IMF Finance & Development (Dec. 2023), https://www.imf.org/en/Publications/fandd/issues/2023/12/POV-Fostering-more-inclusive-democracy-with-AI-Landemore

8 Chris Reed, How AI Can Make Us Better at Arguing, The Conversation (Oct. 27, 2017), https://theconversation.com/how-ai-can-make-us-better-at-arguing-85938.

9 Dave Birss, How to Boost Your Productivity with AI Tools, LinkedIn Learning (July 14, 2023), https://www.linkedin.com/learning/how-to-boost-your-productivity-with-ai-tools/using-ai-to-improve-your-arguments

10 Sudeep Kesh, Shankar Krishnamurthy & Nick Patience, Can Generative AI Create a Productivity Boom?, S&P Global (Jan. 10, 2024), https://www.spglobal.com/en/research-insights/featured/special-editorial/look-forward/can-generative-ai-create-a-productivity-boom

11 Kai-Cheng Yang, Onur Varol, Clayton A. Davis, Emilio Ferrara, Alessandro Flammini & Filippo Menczer, Arming the Public with Artificial Intelligence to Counter Social Bots, HUM. BEHAV. & EMERG. TECH. 2019;e115, available at https://doi.org/10.1002/hbe2.115

12 Yongjun Xu et al., 2021: Xu et al., Artificial Intelligence, 2 The Innovation 100179 (2021).

13 U.S. Department of Education, May 2023: U.S. Dep’t of Educ., AI and the Future, May 2023.

14 Aren Melikyan, An Opportunity and a Threat: The Race for AI, Deutsche Welle Akademie (Sept. 6, 2023), https://p.dw.com/p/4VnUo

15 Lareina Yee, Michael Chui & Alexander Sukharevsky, Generative AI Can Give You “Superpowers,” New McKinsey Research Finds, McKinsey & Co. (July 6, 2023), https://www.mckinsey.com/about-us/new-at-mckinsey-blog/generative-ai-can-give-you-superpowers-new-mckinsey-research-finds

16 Tinker v. Des Moines Indep. Cmty. Sch. Dist., 393 U.S. 503 (1969); Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, 425 U.S. 748 (1976)

17 Davenport v. Wash. Educ. Ass’n, 551 U.S. 177 (2007)

18 Abrams v. United States, 250 U.S. 616, 630 (1919)

19 Hélène Landemore, Dec. 2023: Landemore, Inclusive Democracy, IMF Finance & Dev. (Dec. 2023).

20 Chris Reed, Oct. 27, 2017: Reed, Better at Arguing, The Conversation (Oct. 27, 2017).

21 Lareina Yee, Michael Chui & Alexander Sukharevsky, July 6, 2023: Yee, Chui & Sukharevsky, Generative AI Superpowers, McKinsey & Co. (July 6, 2023).

22 Kovacs v. Cooper, 336 U.S. 77, 81 (1949)

23 Id.

24 Lareina Yee, Michael Chui & Alexander Sukharevsky, July 6, 2023: Yee, Chui & Sukharevsky, Generative AI Superpowers, McKinsey & Co. (July 6, 2023).