There is an experiment that tells us something profound about how we work. In 2012, three researchers from Harvard, Yale and Duke asked a group of people to assemble cardboard boxes, fold origami and build Lego models. They then asked them how much those objects were worth, and the result was both predictable and surprising: participants valued their own amateur creations on a par with those made by experts, and expected others to feel the same. Michael Norton, Daniel Mochon and Dan Ariely called it the ‘IKEA effect’, and it subsequently became very famous (Norton, Mochon & Ariely, 2012).

At first glance it looks like a bias, an error of judgement, but on closer inspection it is something more: it is the mechanism by which the brain constructs meaning. The effort invested creates an emotional bond with the result. To be pedantic, it might be more of a logical fallacy than a fully fledged bias. The more effort we put in, the more our brain needs to justify that investment, and it does so by attributing value to what we have produced. Leon Festinger had already grasped this in 1957 within the framework of cognitive dissonance: when the cost of an action is high, the mind reduces the discomfort by convincing itself that the result was worth it (Festinger, 1957). It is how we turn effort into identity.

The invisible equation

Albert Bandura added a fundamental piece to this picture. In his work on self-efficacy, he demonstrated that successfully completing demanding tasks is the principal means by which people build confidence in their own abilities (Bandura, 1977). It is not so much the result itself that generates self-efficacy as the process of having achieved it through effort. A result obtained without effort does not activate this circuit: the brain registers it, but does not integrate it into its perception of the self.

Effort, attachment, perceived competence, motivation: a self-sustaining cycle that works for as long as there is effort. Remove it, and the cycle breaks.

The age of effort removed

We are living at a moment in history when, for the first time, a technology can produce in a few seconds an apparently intellectual output that would have required hours of work. Generative artificial intelligence writes texts, analyses data, generates code, builds presentations; the quality is often adequate and sometimes surprising.

This article is not against that technology but, paradoxically, in favour of it. It asks a question rarely raised in the debate about how to use it: what happens to the person when effort is removed from work?

The IKEA effect goes into reverse. If perceived value is proportional to the effort invested, then a result obtained without effort is perceived, first and foremost by the person producing it, as less their own or, more precisely, less something they have ownership of. Not because it is objectively worse, but because the mechanism that turns work into meaning has been short-circuited.

Norton and his colleagues revealed a further finding that sheds light on this point: the IKEA effect disappears when people build something and then destroy it, or when they fail to complete the task. Value arises only from the successful completion of work perceived as one's own. When someone else (or something else) does the work, there is completion, but the sense of ownership vanishes.

Three needs, two at risk

In Self-Determination Theory, Edward Deci and Richard Ryan identified three psychological needs that sustain intrinsic motivation: autonomy, competence and relatedness. When these needs are met, people are naturally engaged in what they do. When they are frustrated, the search for motivation shifts elsewhere or dies out (Ryan & Deci, 2000). The systematic removal of effort puts at least two of these needs under pressure.

The need for competence (feeling effective, mastering what we do, growing) is eroded when we delegate the challenging part. The paradox is subtle: we produce more, but feel capable of less. Perceived competence is not built by watching someone solve the problem for us. It is built by solving the problem, however slowly, however many mistakes we make, however imperfectly.

The need for autonomy (feeling ourselves to be the agents of our actions, the authors of our choices) is compromised when the tool stops being a tool and becomes a surrogate. There is a difference between using a technology to explore possibilities and then choosing, and using it to generate a finished product that is handed over as it stands. In the first case we are authors; in the second, intermediaries.

The state we are losing

Mihaly Csikszentmihalyi devoted his life to studying what he called ‘flow’: the state of total absorption that occurs when a person is engaged in an activity that challenges them at the limit of their abilities. Not too easy, not too difficult. Flow, and with it the deepest learning and the most rewarding experience, lives in that space between (Csikszentmihalyi, 1990).

If we systematically remove the challenge from what we do, we are removing the ingredient that activates flow. This goes beyond the loss of emotional engagement: it is a loss of learning, of growth, of that sense of competence that can only be built by tackling something difficult.

It is like a musician using a prerecorded track instead of playing live. The audience hears the music, but the musician can no longer feel their own fingers.

The inflation of meaning

There is also a collective dimension that extends beyond the individual. When everyone can produce the same kind of result with the same tool, that result loses value. Perceived value, even before economic value. What yesterday required rare expertise is today accessible to anyone. Code, text, analysis, presentations: products that defined the professional identity of their creators are becoming indistinguishable from those generated by anyone else with access to the same technology.

The phenomenon is broader than it seems, and it is not just about the labour market. It is about the stories people tell themselves about who they are. If what I once produced through effort is now produced by a machine, what remains of my definition of myself through work?

Value shifts from the output to the thinking behind it, to the capacity for judgement, to the depth of our knowledge of our field, to the quality of the questions we ask ourselves before producing anything. But these abilities are not developed by delegating; they are developed by exercising them.

The surface that looks like depth

There is a further aspect concerning those who receive the results of our work. Generative artificial intelligence produces outputs with an insidious characteristic: they look complete. The structure is right, the vocabulary appropriate, the logical connectives in the right places. Everything is ready to lead a superficial reader to an easy conclusion: yes, there is quality here.

But the quality of content should lie (and does lie) in the accuracy of the information, the depth of the analysis, the ability to distinguish what is relevant from what is merely plausible. AI produces statistically plausible outputs, but not necessarily accurate ones. This difference is invisible to those reading in haste, and over time it accumulates: content begins to look alike, analyses remain on the surface and conclusions become predictable.

Those who do not check, investigate further or question what they receive from a machine contribute, without realising it, to a collective impoverishment of the depth with which we address the things that matter.

Effort as a doorway

The solution is not to give up the tool: that would be like rejecting digital technology when it arrived; some people did, and it did not go well. The solution is to change the kind of relationship we have with it.

Artificial intelligence delivers its greatest value when used as an interlocutor, not a substitute. This means a change of approach: do not ask the machine to produce the result, but ask it to challenge the result we have produced ourselves. Do not delegate the research, but use it to check and enrich the research we have done.

Sources are a concrete example. An AI produces quotations and references with apparent confidence; checking them is a process that requires effort. And it is precisely that effort that opens unexpected doors. One source leads to another, one piece of data contradicts another, an unfamiliar author offers a perspective we had not considered. Verification is not a cost but the moment when we learn something new. It is the challenge that activates flow, the completion that builds self-efficacy, the effort that generates attachment to the result.

Using technology as a sparring partner keeps the challenge where it is needed: in critical thinking, evaluation and judgement. It keeps intact the mechanism that turns work into meaning.

The question that remains

People develop through challenges. Research on flow tells us this, work on self-efficacy tells us this, Self-Determination Theory tells us this, and so does the experience of anyone who has ever learnt something difficult.

The reward lies not in the result but in the journey.

If we delegate the journey, the result is there but the growth is not. Without growth, what we do stops telling us who we are.

Artificial intelligence is a tool that amplifies, but an amplifier worthy of the name strengthens a signal that already exists. If the signal is strong, the amplification is extraordinary; if the signal weakens because we have stopped exercising it, the amplifier creates nothing.
Amplifying emptiness or background noise rarely creates anything pleasant to live with.

We are the signal; keeping it alive is a choice no technology can make for us.

References

Norton, Michael I.; Mochon, Daniel; Ariely, Dan, “The IKEA Effect: When Labor Leads to Love”, in Journal of Consumer Psychology, vol. 22, Hoboken, 2012, pp. 453–460.

Festinger, Leon, A Theory of Cognitive Dissonance, Stanford, 1957.

Bandura, Albert, “Self-Efficacy: Toward a Unifying Theory of Behavioral Change”, in Psychological Review, vol. 84, Washington, 1977, pp. 191–215.

Ryan, Richard M.; Deci, Edward L., “Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being”, in American Psychologist, vol. 55, Washington, 2000, pp. 68–78.

Csikszentmihalyi, Mihaly, Flow: The Psychology of Optimal Experience, New York, 1990.

Draxler, Fiona; Werner, Anna; Lehmann, Florian; Hoppe, Matthias; Schmidt, Albrecht; Buschek, Daniel; Welsch, Robin, “The AI Ghostwriter Effect: When Users Do Not Perceive Ownership of AI-Generated Text but Self-Declare as Authors”, in ACM Transactions on Computer-Human Interaction, vol. 31, New York, 2024, pp. 1–40.