By: Ahren Lahvis
ChatGPT has set the world ablaze. The publicly available and free-to-use chatbot is an application programming interface (API) that generates responses to language requests through artificial intelligence (AI), and processes millions of such requests per day.[1] Released for public access in November 2022, ChatGPT can, upon request, produce jokes, TV episodes, music, and computer code.[2] Students now use it to write papers, businesses use it to create promotional materials, and lawyers use it to draft legal briefs.[3]
ChatGPT was developed by OpenAI, an artificial intelligence research company whose mission is to “ensure that artificial general intelligence benefits all of humanity.”[4] As is the case with groundbreaking technology, however, the benefits conferred on society are often accompanied by serious shortcomings.[5] One such concern for ChatGPT—in addition to worries about plagiarism and malware creation capacity—is its potential negative impact on the workforce. Let ChatGPT itself summarize the issue:
Artificial Intelligence (AI) has the potential to negatively impact employment and labor by automating certain tasks and jobs, leading to job displacement and unemployment. This could particularly affect low-skilled and low-paying jobs, as well as workers in industries that are highly susceptible to automation. Additionally, the increasing use of AI in the workplace may lead to a lack of job opportunities for workers with less technical skills, exacerbating economic inequality. Furthermore, AI-powered machines and software may displace human workers and could lead to wage stagnation as the same work can be done by machines with lower costs.[6]
In 1933, John Maynard Keynes called unemployment produced by major technological change “technological unemployment,” which is caused by “discovery of means of economising the use of labour outrunning the pace at which we can find new uses for labour.”[7] This phase, however, is subject to a “temporary phase of maladjustment.”[8] The Economist clarifies, “technological progress has never previously failed to generate new employment opportunities,” citing unemployment recoveries in the wakes of advances in computational and agricultural technology.[9] Artificial intelligence, however, is a potentially different beast. Agricultural technologies and factory equipment impacted largely manual tasks and computers impacted math-based white-collar roles like accountants and mathematicians. Artificial intelligence, however, threatens a different kind of worker—namely, the journalist, paralegal, or songwriter, who relies on cognitive, non-artificial intelligence, a subset of the workforce that is historically considered immune to the effects of automation.[10]
ChatGPT, like its technological predecessors, may also threaten organized labor, at least in the short term. Automation and technological progress have long been thought to negatively impact unionization by changing the nature of the work, reducing workforce solidarity, and impeding strike efforts.[11] For company management seeking to sidestep a strike—the “ultimate weapon of free trade unions”[12]—automation allows a company to run a factory on fewer workers who are outside of the bargaining unit, such as supervisors.[13] Technological progress also generates wealth and often provides workers with increased living standards, which can mitigate worker interest in joining a union.[14] Predicting AI’s effect on white collar employment is nigh-on impossible, but the past is a less-than-optimistic precedent.
Labor in general is also on the decline, having decreased across the globe since the turn of the twenty-first century. The Organisation for Economic Co-Operation and Development (OECD), an international organization collecting economic data from thirty-eight countries, shows that union density dropped by over five percent from 2000 to 2019.[15] In Australia, the percentage of the workforce that was unionized dropped from 24.9% in 2000 to 13.7% in 2018.[16] During the same period, union density in Finland dropped from 74.5% to 60%.[17] And in the United States, the number dropped from 12.9% in 2000, to 9.9% in 2019.[18]
Artificial intelligence’s effect on organized labor in the United States may also be compounded by a Supreme Court keen on stripping down union protections. The Court this session seems poised to affirm employers’ ability to bring a tort claim against unions that damage property during strikes.[19] In recent years, the Court’s conservative majority has also hamstrung union operations by holding that public-sector unions cannot collect agency fees from non-consenting employees,[20] and that regulations allowing union organizing on farms constitute unconstitutional physical takings.[21]
The United States is not alone in rolling back worker protections. In the United Kingdom, Parliament is seeking to enact a law that would require minimum service levels during strikes, in an effort “to ensure that striking workers don't put the public’s lives at risk.”[22] Doing so will likely have a chilling effect on strikes, which are effective tools because of the leverage gained by withholding services completely. In France, President Macron’s proposal to increase the retirement age from sixty-two to sixty-four have caused widespread protests and strikes.[23] The South Korean government is currently considering a law that will allow workers to exchange overtime hours for time off and raise the maximum allowable overtime hours from fifty-two to sixty-nine.[24] Workers are concerned that the law will negatively impact the economic freedom of women and experts are skeptical about the plan’s ability to meet its goal of reducing overall work.[25]
Enter ChatGPT. The API does not pose a threat to employment in manual labor industries such as agriculture, food service, or construction, for example, but it has serious potential to disrupt workforces in white collar industries. ChatGPT’s effects on the labor force are impossible to predict, but the widespread availability of the powerful API comes at the risk of job displacement when, in the United States and around the world, worker protections and job security are increasingly in peril. The current concoction of legal rollbacks, AI development, and financial instability may spell danger for workers’ rights and organized labor, creating a problem that even ChatGPT cannot compute.
Ahren Lahvis is a Staff Editor at CICLR
[1] About, OpenAI, https://openai.com/about/ (last visited Jan. 21, 2023), [https://perma.cc/2ZMU-NTWD]. [2] Rebecca Heilweil, AI is Finally Good at Stuff, and That’s a Problem, Vox (Dec. 7, 2022), https://www.vox.com/recode/2022/12/7/23498694/ai-artificial-intelligence-chat-gpt-openai, [https://perma.cc/ZFY4-HHDB]. [3] Annie Lowrey, How ChatGPT Will Destabilize White-Collar Work, The Atlantic (Jan. 20, 2023, 6:00 AM), https://www.theatlantic.com/ideas/archive/2023/01/chatgpt-ai-economy-automation-jobs/672767/ [https://perma.cc/W9YH-XPPL]. [4] About, supra note 1. [5] See, e.g., the discovery of nuclear fission. Michael Madsen, Pioneering Nuclear Science: The Discovery of Nuclear Fission, Int’l Atomic Energy Agency (Dec. 20, 2013), https://www.iaea.org/newscenter/news/pioneering-nuclear-science-discovery-nuclear-fission [https://perma.cc/Y7ZD-N3BP]. [6] ChatGPT, https://chat.openai.com/chat (last visited Jan. 21, 2023). This paragraph was created in response to the following prompt: “Write a short paragraph about how artificial intelligence may negatively impact employment and labor.” [7] John Maynard Keynes, Essays in Persuasion 325 (Palgrave Macmillan 3d ed. 2010) (1931). [8] Id. [9] The Onrushing Wave, The Economist (Jan. 18, 2014), https://www.economist.com/briefing/2014/01/18/the-onrushing-wave [https://perma.cc/6RXG-HJ2F]. [10] Lowrey, supra note 3. [11] James L. Stern, Automation—End or a New Day in Unionism?, 350 Annals Am. Acad. Pol. & Soc. Sci. 25, 26-28 (1963). [12] Id. at 27. [13] Id. [14] Id. at 28. [15] Trade Union Dataset, Organisation for Econ. Co-Operation & Dev., https://stats.oecd.org/Index.aspx?DataSetCode=TUD (last visited Jan. 21, 2023) [https://perma.cc/B2AR-7LEP]. [16] Id. [17] Id. [18] Id. [19] John Kruzel, U.S. Supreme Court Considers Narrowing Federal Protections for Unions, Reuters (Jan. 10. 2023, 2:00 PM) https://www.reuters.com/legal/us-supreme-court-considers-narrowing-federal-protections-unions-2023-01-10/ [https://perma.cc/ZP2H-WKSW]. [20] Janus v. AFSCME, 138 S. Ct. 2448, 2459-60 (2018). [21] Cedar Point Nursery v. Hassid, 141 S. Ct. 2063, 2079-80 (2021). [22] Faisal Islam, Workers’ Rights Agency Denies Backing New UK Strike Law, BBC (Jan. 18, 2023), https://www.bbc.com/news/64313013 [https://perma.cc/LH7T-2DAB]. [23] Anthony Paone & Benoit Tessier, Trade Unions Call for More Strikes over Macron’s Pension Reform, Reuters (Jan. 19, 2023, 8:24 PM), https://www.reuters.com/world/europe/nationwide-strike-france-against-macrons-pension-reform-2023-01-19/ [https://perma.cc/4MZU-69EM]. [24] Hyunsu Kim, As South Korean Government Proposes Flexible Overtime Rules, Some Fear Workers May Suffer, Reuters (Jan. 20, 2023, 3:05 AM) https://www.reuters.com/world/asia-pacific/south-korean-government-proposes-flexible-overtime-rules-some-fear-workers-may-2023-01-19/ [https://perma.cc/USV2-K3L3]. [25] Id.
Comments