Prepare students for ASI

ASI is a theoretical system that will function on an entirely different scale from current AI, operating with a capacity so advanced that it is as far beyond humans as humans are beyond chimpanzees.

It will affect the way we live and work. Geoffrey Hinton, a k a “The Godfather of AI,” is known for pioneering work on neural networks and deep learning. He has spoken publicly about the risks of advanced AI systems. In 2023, he left Google to speak more freely about his concerns, including the social impact of powerful AI technologies. He has warned that AI should be handled with care and not treated as a purely beneficial trend. When asked what job people should train for, Hinton’s advice was blunt: “Become a plumber.”

The point is that many forms of work, especially those requiring hands-on physical skills, are harder to automate than desk-based tasks. Plumbers, electricians, heating engineers, and maintenance workers will still be needed, and health and care work still depends on human judgement and human support.

With the advent of AI/ASI, jobs will not just be about producing results, but making accurate judgements and being accountable for decisions. This should reshape how—and what—universities teach.

Some university courses still train students for tasks that AI can already do more cheaply and quickly. For example, IT courses frequently emphasise writing code, building websites and generating reports. Business Studies often spends too much time on producing charts, drafting marketing copy, or generating routine paperwork. If students are trained primarily to generate results rather than to verify their accuracy, their core skills will not compete with AI/ASI. Routine work is already being replaced by AI automation: driverless or driver-minimal taxi services have begun operating in parts of the world, and supermarket checkout jobs are shrinking as customers increasingly scan and check out themselves.

AI changes the value of human labour: it shifts the emphasis from producing results to checking, approving and taking responsibility for outcomes. Human input is needed to verify the accuracy of AI results and decide whether to accept them. In many subjects — not just in IT and Business Studies — universities need to place more focus on decision-making and verification skills.

There are several ways universities should consider approaching the problem. First, they should train students to verify and audit outputs, not just generate them. Students should be assessed on their ability to ask basic questions of AI/ASI: Is this correct? How do we know? What could go wrong? What evidence would we need before trusting it? Assessment should reward error-spotting and correction, because that is a skill AI/ASI cannot so easily replace. Second, courses should develop decision-making capabilities. Workplace problems are complex and ambiguous. People disagree and plans will have to change rapidly to meet the ASI revolution. Students who only learn how to finish predefined, routine tasks can struggle when they must decide quickly under pressure. Projects, work placements, and apprenticeships can help students practice making judgement calls in realistic conditions.

Third, knowledge about AI/ASI should be taught as part of everyday work across all degrees, not treated as a separate optional topic. Since students already use basic AI tools on their own, the curriculum must teach the rules: when to rely on AI/ASI, when not to, how to challenge outputs and above all, how to stay in control.

Finally, universities should avoid curricula that primarily train students for work that AI/ASI can or will do more efficiently. The aim should be to build the human skills AI/ASI can’t easily replace, on top of whatever it can. If those changes are made, graduates leave with skills that remain valuable as more routine tasks become automated.

If universities do not act, the mismatch between what students are trained for and what the job market demands will grow, including for students graduating soon. ASI is not a distant future and it will transform how we live and work.