By Terence Tse
AI may automate certain tasks, but it rarely eliminates entire jobs. The real cost of displacement goes beyond lost income, threatening the purpose, identity, and dignity that work provides.
The public debate surrounding AI is often dominated by a single, catastrophic headline: that machine intelligence will inevitably take away all our jobs. Tech executives routinely warn of impending mass joblessness. Not a day goes by without a company cutting positions to realise cost savings from using AI. Yet this prevailing narrative is profoundly misleading.
Of course, AI has taken over some human jobs; just ask proofreaders and translators. But we must recognise and dismantle two misconceptions. First, automation is not an indiscriminate job-destroyer. Second, the resulting losses are often not purely financial.
The displacement argument is nuanced
The claim that AI will wipe out employment rests on a fundamental error: confusing an individual task with an entire job. When economic analyses suggest that a job is “exposed” to AI, they don’t mean the job as a whole is under threat. Instead, they describe how algorithms can affect specific, discrete activities, both positively and negatively, not how much the position itself can be eliminated.
When a job contains ten tasks and AI replaces one, there are still nine tasks for humans to do.
Many jobs have long been argued to be “messy”: workers must handle multiple streams of information, respond to a dynamic environment, work with unclear or shifting goals, and multitask. Dealing with such unstructured workflows is, at least for now, beyond AI’s capabilities. To be sure, algorithms can take over a single structured, definable task. Yet tasks in a job don’t usually operate in isolation; they are complementary and interdependent. So, when software handles routine analysis, drafting, or data organisation, people don’t just disappear from the workplace. Instead, it creates what is called the “focus effect”. By freeing up time from repetitive work, employees can spend more energy on complex problem-solving, building client relationships, or using their judgement. As a result, they often become much more productive in these areas, which can lead to higher value and better pay.
Ironically, this can change companies’ incentives to deploy AI. Consider using an AI that is cheap but worse than humans at a task. The resulting value created by the job would fall. In this case, saving money with mediocre AI can backfire because its weaknesses drag down all the good work done in other tasks. By contrast, if the AI helps humans do other non-automated tasks better, it can improve the entire job output. AI benefits therefore depend on how tasks make up a job.
Now think about the so-called “job dimensionality”, ie. the number of tasks in the job itself. This matters because a job with only a few tasks is likely to be fully displaced. When a job contains ten tasks and AI replaces one, there are still nine tasks for humans to do. But when a job has only one task, the job is gone. From a company’s perspective, there is much more incentive to invest in AI for the second case because it is much easier to realise the cost savings involved.
It is possible that demand for human-based output is set to rise. As the rich get richer, they want more from other humans, not less. They want things that aren’t commodities in the conventional sense: the social aspects of products, such as relationships, status and exclusivity, start to matter far more. Their spending shifts towards goods and services where the human element, the experience or the social meaning counts. They want clothing with a story, food with history, and personal trainers who can tailor their approach to their needs. Demand for these qualities brings the human element back into production, and with it, jobs. Value is therefore like to migrates into an expanding “relational sector”—spanning education, healthcare, and artisanal crafts—where the human presence is precisely what consumers demand.[1]
Money is only a part of our lives
The current discourse often focuses on ensuring the economic survival of those who have lost their jobs to automation. Anthropic claims it will put 5% of its shares into a public-benefit fund to help. Elon Musk pledged to give one SpaceX share to every child in the US. However, the belief that work is merely about chasing a monthly pay cheque is wrong. Decades of research have shown that work provides more than money. The social psychologist Marie Jahoda identifies five indispensable psychological anchors that paid work provides: a daily time structure, routine social contacts outside the immediate family, a connection to a collective and a shared purpose, social status afforded by having a job, and nudges toward purposeful activity. These may explain why Swedish lottery winners have been found to reduce their hours or take more time off rather than quit entirely.[2]
Losing a job means more than just losing a pay cheque. It can also bring serious psychological and social challenges. A recent study showed that working-age adults who leave their jobs before retirement may experience faster cognitive decline.[3] Job loss can even shorten lives. In Pennsylvania, men who lost their jobs in mass layoffs had death rates 50 to 100% higher than expected in the year after losing work. Even 20 years later, their risk was still 10 to 15% higher.[4]
there is much more incentive to invest in AI for the second case because it is much easier to realise the cost savings involved.
Joblessness also makes people less satisfied with life, and this feeling can last for years. In a 15-year study of German workers, most people never fully got back to their previous level of life satisfaction, even after finding new jobs.[5] On the bright side, even a small amount of volunteering can improve life satisfaction, provide a sense of purpose, and support personal growth.[6] Another study found that people who are unemployed for a long time only feel much happier when they finally reach retirement age. Their daily routines don’t change, but they move from being part of an unwanted group to a respected one.[7] Having a paid job gives people a sense of dignity.
The connection between AI and job loss isn’t simple. It’s more complex, and we need to look at the details to understand what might happen next. AI takes over tasks, not whole jobs, and when it helps people do the rest of their work better, it can make that work more valuable. Treating displacement purely as an income gap overlooks what work provides: structure, community, status and dignity. A cheque can replace a salary, but not a reason to get out of bed in the morning.
About the Author
Terence Tse is Professor of Finance at Hult International Business School and co-founder at the AI Native Foundation. He is also co-founder and Executive Director of Nexus FrontierTech.
