Chapter 4, Part 4 — What if Further Developments of AI are Banned

Chapter 4, Part 4 — Applying AI to Education

Scenario #1: Further Developments of AI are Banned

This category describes a future whereby the US, EU, and China’s governments become fearful of the existential threat of AI and decide to ban any future development of the technology. While this idea may seem too far outside of the realm of possibility, many well-respected AI experts are already deeply concerned about the possibility that AI could get banned in the future, including Max Tegmark, Nick Bostrom, Yuval Noah Harari, Elon Musk, Steve Wozniak, Yoshua Bengio, Stuart Russell, Michael Osborne, Daron Acemoglu, and Eliezer Yudkowsky (Hamilton, Wiliam, & Hattie, 2023).

Authors Hamilton et al. argue that a ban on future development is not without precedent, citing how countries have come together in the past to ban future development of technology (Hamilton et al., 2023). They cited examples including banning human cloning in 2005, the Treaty on the Non-Proliferation of Nuclear Weapons (1970), and lesser-known examples of global bans including the Chemical Weapons Convention; the Biological Weapons Convention; and the Convention on the Prohibition of the Development, Production, and Stockpiling of Bacteriological (Biological) and Toxin Weapons and on their Destruction (BTWC) to name a few (Hamilton et al., 2023). However, the researchers warn that even if future developments of AI technology were banned, the relatively cheap and conspicuous nature of developing and training AI algorithms would make it hard to regulate this ban (Hamilton et al., 2023).

Heightening their concern is that AI relies heavily on advanced semiconductor chips, mainly produced by a handful of companies. As Chris Miller notes in his book Chip War, most semiconductor chips utilize software from only three U.S.-based companies—Cadence, Synopsys, or Mentor (owned by Germany's Siemens but based in Oregon) (Miller, 2022). Furthermore, advanced processors require EUV lithography machines, exclusively made by the Netherlands' ASML, which in turn depends on its San Diego subsidiary, Cymer, for essential light sources. This means that it would also be theoretically possible for governments to temporarily put the production of these chips on pause while negotiating the terms of this AI ban (Miller, 2022).

Under this scenario, education may continue to mostly look and function similarly to what it looks like today. Even as AI technology finds its way into classrooms, it would likely evolve in a more iterative fashion. Under this scenario, existing AI-enabled educational technology, with some inevitable advancements, does not really disrupt the status quo, much in the same way computer technology was eventually absorbed into the daily work of students without really changing the fundamental “factory model” referenced earlier. If the ban is successful, a more innovative educational model being pioneered by the Alpha School could be possible to break the old educational models and perhaps even make a difference in how millions of students learn, are instructed, and assessed.

Yet, the authors suggest that the time it takes to create, negotiate, and implement a treaty that essentially bans, or considerably slows down, future development of AI won’t be able to outpace the speedy growth of AI we are already witnessing (Hamilton et al., 2023). They believe that by the time AI is banned it could be too late, leaving the other three option,s they delineate the likelier outcomes to consider. These are explored in the next sections.

References

Hamilton, A., Wiliam, D., & Hattie, J. (2023). The future of AI in education: 13 things we can do to minimize the damage [Working paper]. Cognition Education. https://cognitioneducation.com/news/ai-in-education/

Miller, C. (2022). Chip war: The fight for the world's most critical technology. Scribner.