Chapter 4, part 5 - What if AI Remains Unregulated

Chapter 4, Part 5 — Applying AI to Education

Scenario #2: Human-AI Partnerships

The Human-AI partnership imagines a future whereby AI remains mostly unregulated and eventually drives a shift in the work dynamic from humans leveraging AI to enhance their workflows to AI fully managing all job aspects. This includes conducting analyses, formulating plans, executing tasks, and handling interactions with humans (Hamilton, Wiliam, & Hattie, 2023). Research on labor-market analyses by Goldman Sachs, McKinsey, and OpenAI suggests that current AI technology poses a risk of automation to as many as 300 million jobs globally, which, surprisingly, focuses mostly on higher-paid knowledge work positions (Hamilton et al., 2023).

The reports the researchers analyzed also mentioned that AI will help to catalyze innovation that will also result in many new jobs, products, and services that we can’t yet imagine. AI-driven automation might replace some existing jobs, but it could also give rise to entirely new and currently unimaginable forms of employment (Hamilton et al., 2023). However, the researchers disagree with the idea that humans will always be able to fill these new roles like they always have (Hamilton et al., 2023).

Instead, the researchers describe two scenarios that may result from human-AI partnerships, each with implications for education: The first scenario is where all human workers, teachers included, have digital assistants that support their work but are evolved enough that they can do everything humans can do. The other scenario is driven by an economic downturn whereby large global companies gradually cut down their highly educated workforce and replace their roles with AI (Hamilton et al., 2023).

While this idea may seem far-fetched, the researchers argue that these principles are already part of our modern economy. The researchers point to the London Underground, where trains have driven themselves since 1968, yet still pay a person to sit at the front (Hamilton et al., 2023). They also point to airplane pilots as well,l where the autopilot is engaged for roughly 90% of the flight, and has the capacity for automated landings, just requiring the pilot to help take off. They also note that fully automated airplanes currently exist, but they are just not yet in commercial use (Hamilton et al., 2023).

David Graeber's 2019 book titled Bullshit Jobs argues that already 50% of existing jobs are utterly pointless (Graeber, 2019). Graeber suggests that this is a product of managers who derive power and status from the size of the departments they lead, although much of the work adds little to the effectiveness of the organization (Graeber, 2019). The researchers reference this idea to contrast it with a notion of “Fake Work” whereby some aspect of work is deemed “important”, allowing even the most tenuous jobs to enable the worker to say: ‘I made a difference today’. But considering the longer-term future where AI can do almost everything better than we can, the feeling of empowerment could very likely become “I slowed things down today: I got in the way of the machine?” (Hamilton et al., 2023).

In situations where humans and machines work together, schools might evolve into a place that teaches the basics of civic life, provides spaces for social interaction, and teaches how to work with AI. However, as AI takes over most of the essential tasks, there could be little reason to learn many of the skills currently emphasized by the education system, especially in higher education (Hamilton et al., 2023).

This looming technical reality will have far-reaching implications for education. Firstly, students may need to shift their focus from learning specific tasks to mastering how to manage machines that perform those tasks more effectively. Secondly, there lies the significant challenge of motivating students to engage in learning. Why should students strive to learn if machines can perform nearly all tasks better than humans? This scenario presents substantial economic and social challenges with extensive consequences for education, all of which merit further study and are beyond the scope of this paper.

References

Graeber, D. (2019). Bullshit jobs: A theory. Simon & Schuster.

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/