SEVERAL people (myself included) have written about the harmful impact of LLM (large language model) chatbots on basic learning processes in educational institutions. Learning is not merely accessing content, it is thinking with and through content. This is why those who claim that schooling (or processual education at any stage) can be easily replaced with an LLM subscription and internet access are selling a dangerous fiction.
Over a longer period, I am more optimistic that institutionalised learning will stay relevant, perhaps even without excessive reliance on technology. Already we see schools across several high-achieving educational contexts experimenting with tech-free classes and reporting improved student learning performance. Schools and universities themselves will continue to act as key sites of human socialisation, which should remain relevant in a world increasingly fragmented, insular and diffused.
What is far less clear, and where the possibility of rupture and crisis is more plausible, is the labour market itself. For just over half a decade, I’ve been teaching a class on markets and the roles they play across different societies in history. The premise of this class is to familiarise students with how our current reality, where needs and wants are mostly satisfied through market-based exchange (money swapped in exchange for commodities) is a historically contingent outcome. The other purpose is to show how this contingency itself is subject to considerable variation — many societies differ in the extent to which they allow markets to shape the lives, livelihoods, and afterlives of their citizens.
As one would expect, the issue of AI has dominated quite a few of my in-class conversations these past few years. Students are often curious about the global future that mass adoption of AI will herald. Many are concerned with their own personal futures, with regards to skills, employment, and career trajectories. There is a vague, though pervasive, concern that the future is uncertain and that received wisdom on what’s needed to live a materially successful life simply won’t be applicable in the years ahead.
Curious about what the adoption of AI will herald, many students are concerned with their own personal futures.
These concerns are valid. Young people the world over remain disillusioned with the abrogation of political and social rights, mass violence, and mounting economic inequality. Increased uncertainty about their fate in a rapidly changing labour market exacerbates their disillusionment even further.
One set of mainstream views on AI’s impact on the labour market tends to view it as something that will enhance labour productivity to a very high degree, i.e., make humans do certain tasks much more efficiently. This view draws on an understanding of employment as ‘task bundles’. Technology can thus help with some tasks and not with others. Human capability in specific jobs can then be enhanced because of increased task efficiency. The prediction follows that productivity gains from this enhanced efficiency will lead to greater and/or more sustained growth. Yes, some jobs will be made redundant but just like cash-tellers were readjusted into the economy after the invention of the ATM, everyone will (eventually) land on their feet.
There are a couple of issues with this mainstream view, even if we take the dramatic productivity enhancement value of AI as a given (which though promised isn’t very visible as yet). Most obvious of these is what happens during the ‘adjustment period’ when a subset of workers is displaced and made redundant. These losses may be concentrated in some sectors, or even in some geographic locations. Countries dependent on business process outsourcing or IT exports, for example, may face a bigger crisis than those trading in other types of goods and services.
Another issue is that the jump from productivity enhancement to broad-based economic growth is a bit of a mystery. What is the pathway to this growth? As Malcolm Harris asks, what happens if you’re able to code faster, or do document filing more efficiently? Companies save on time and resources, and probably on labour costs, which increases their profit margins but does that necessarily increase output? Will economies produce more stuff or provide more services just because it’s faster to do it? The former niche Marxian view that the world economy already suffers from a problem of overcapacity is now something many mainstream economists openly admit. So again, a worthy question to pose is if LLMs improve productivity and remove labour-based bottlenecks, how does that necessarily deliver growth?
And this brings us to the main issue with the current system in which AI is being touted as the centrepiece of the future. In the most techno-futurist scenario, often peddled by Silicon Valley entrepreneurs and their acolytes, humans will largely be made redundant by a combination of AI and robotics across all industries. This combination means that there is no human labour contribution to total output and thus no share for human labour in total income. Everything will flow back into the hands of those who own capital (ie, AI and robots). This configuration of income flowing on the basis of ownership is how most of the world is currently organised.
Think about what this scenario means for societal functioning as a whole. If the only income available to be spent is in the hands of those who control technology (ie, productive resources), what happens to the basic needs of those who simply used to sell their labour power? In the absence of any profit to be made selling to income-less humans, will these populations be left to waste?
I don’t think such a future will arrive, though it is telling that it’s the one most frequently pushed by the tech world. If the supposed productivity gains are ever realised, there will be fierce political conflict over the distribution of those returns. But it is quite clear, and somewhat ironic, that the supposed technology revolution unleashed by capitalism will also require us to rethink the centrality that markets and private ownership of resources play in our lives.
The writer teaches sociology and politics at Lums.
X: @umairjav
Published in Dawn, February 2nd, 2026
