Why AI Won’t Transform Work as Fast as You Think

Amid the widespread apocalyptic predictions of the impact of AI on the future of work, we would do well to remember what has become known as Roy Amara’s law: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”

Analyses of an imminent collapse in employment or radical automation often ignore the reality that we don’t adopt new technologies in a linear or homogeneous manner. Recent studies such as “The Productivity Puzzle: AI Technology Adoption and the Workforce” show that the so-called “productivity explosion” does not yet exist: instead, there’s a lot of talk, along with experimentation and a gap between potential and reality.

The ongoing narrative about AI is little short of apocalyptic: jobs destroyed, corporations without employees, robots doing our jobs. But if we look carefully at the most recent evidence, the picture is very different: it is more moderate, slower, more similar to evolution than to revolution. And in that sense, the most likely result is that in ten years we will be working in a very similar way to today: more AI, no doubt, but the essence of work won’t be radically different.

Articles along the lines of The Wall Street Journal’s “How the internet rewired work — and what that tells us about AI’s likely impact” explain how when the web began to transform everything, many imagined the immediate end of the office. face-to-face work and regular employment. There was talk of economies without physical employment, of wholesale teleworking, new jobs that were going to emerge en masse. However, although the internet changed many things, much of the work remained the same: meetings, hierarchies, formal routines. The essential remained. We are in an equivalent phase with AI: high expectations, experimental adoptions, modest results. MIT Sloan’s recent research points out that when companies adopt AI, they often experience an initial drop in productivity before stronger growth takes shape. This brings us to a “J”-shaped curve: first stagnation, then take-off.

This idea that the impact of a technology takes time to manifest itself has profound implications, which we have seen before: if companies, regulators or individuals expect an immediate reorganization of work, they may become disillusioned or act rashly. But if we accept that work will slowly evolve based another logic, that of complementarity, the reorganization of flows and the integration of humans and machines, then we can prepare ourselves better. Recent studies have found that much of the use of artificial intelligence in the workplace is leading to more monitoring, evaluation and correction tasks than a simple replacement of labor: AI as an assistant, not a replacement.

As a result, jobs don’t disappear: their nature changes. Routine tasks may be reduced, but new functions will be created that manage, monitor or complement AI. And that transition does not happen from one month to the next. For example, recent surveys show that a high percentage of workers feel that AI has increased their workload or that it hasn’t generated the expected productivity improvement. We are not facing an era without jobs, but one of adjustment.

From a European perspective, we need to make a commitment to permanent training, adaptability, hybrid skills that combine human and technological capacity. Because if the focus of work continues to be the person, even if it is now supported by AI, we must ensure that the transition is managed fairly, that there are no new gaps and that the value of work remains valid, albeit different.

It’s a mistake to think that AI is going to “liquidate” jobs and that in ten years we will all be unemployed or subcontracted by algorithms. But we must also avoid falling into the naïve optimism that nothing is going to change. It will do so, but continuously, unprecedented in its accumulation, invisible in its pace. The technology that transformed work did not do so in a year or in a decade but in several. And even today we continue to see the effects of the internet, the cloud, the mobile phone and algorithms in our working day.

I believe we should be prepared for slower change than is being predicted, and that we use that time to manage the transition well. Because if we put all our expectations in an immediate leap, we likely ignore the intermediate steps that really transform work: the reorganization of processes, the design of tasks focused on humans and machines, the training in new skills. When that’s done right, the impact can be considerable… but it will take years.

In short, AI will not be the broom that sweeps away work as we know it, but the lens that will allow us to see it with other nuances. Ten years from now, we’ll probably be in a recognizable job, perhaps with more AI assistants, new supervisory tasks, or more automation, but with the same basic structure: people who work, organizations that organize and processes that produce. And that’s not a failure of AI, it’s simply the classic pattern of technology: a slow start followed by cumulative, deep change over the long haul.

If we accept that, we will be more prepared to lead the transformation rather than simply being swept up in it.

(En español, aquí)

—

This post was previously published on MEDIUM.COM.

—