AI is making programmers more productive than ever — but could also be eliminating junior roles in their industry, writes Jason Walsh.
For a decade, ’learn to code’ was the proposed answer to every anxiety about the future of work. But what if it turns out the coders aren’t safe either?
The CV that lands on a recruiter’s desk now looks much the same as it did five years ago. What has changed is what happens next: increasingly, the junior developer role it was written for no longer exists.
One thing we know about artificial intelligence (AI) is that it is productive. AI augmentation has arisen as the key promise of the technology, supercharging the productivity of everyone from programmers to information security specialists.
Beneath this, though, there is a clear worry, with many ruminating on a dark question: what if augmentation is a red herring? The irony is not lost on anyone: the workers most exposed to AI displacement are the same ones who built it, deploy it and, increasingly, rely on it to do their jobs. For Ireland’s tech sector, long insulated by the presence of US multinationals and a skills shortage that kept salaries high and CVs moving, the question is no longer whether AI will change the nature of work, but how fast, and who will be left holding the bill when it does.
A recent paper by researcher Yann Harel published in the scholarly journal AI and Society said that mental labourers, including software developers, data scientists, media professionals, paralegals and finance workers, are at risk.
The paper, entitled ‘AI, mental and physical labour, and a just policy framework’, argues that “considering that this technology is only a few years old and assuming it will keep developing, the influence of AI on the labour market is likely to increase, and we may witness an unprecedented wave of displacement of mental labourers from a wide variety of occupations”.
Citing another paper, it notes a 17 to 21 per cent decrease in the number of job posts.
THE BOTTOM RUNG
Martin Duffy, managing partner at Hadfield Green, which specialises in senior and executive recruitment, says the impact of AI is already being felt locally.
“One thing we’ve noticed is a sharp fall in graduate hiring, especially in tech; routine entry-level roles are being cut back on. They have AI augmented staff already.”
Recruitment has not ceased so much as become more selective, he says. In other words, so-called ‘milk round’ recruiting may not have a future and, in tech specifically, AI is already replacing coding and data tasks.
“Graduate positions are still being offered but employers are assessing the value of hiring large cohorts of graduates,” Duffy says. “Speaking to CTOs, they’re just not hiring like they used to. A principal software engineer will use an AI tool to do what a junior employee would have done in the past.”
Peter Rose, group chief information officer at digital transformation consultancy TEKenable, offers a different framework for AI: AI employees.
Readers could be forgiven a sense of ontological vertigo when confronted with such an idea: partly a metaphor and partly a method for accounting for AI spend, AI employees are agents designed to replace or augment specific human roles.
“One is the assistant level, which is copilot. Level two is agentic but driven by a human being saying, ‘do this’. Level three is an agent making semi-autonomous decisions, but consulting with its human when it needs to.
“You only let something have level three when it has proven itself at doing something and only within certain boundaries,” Rose says.
However, the question remains about what technology specialists should do. One thing to note is that the disruption will not be evenly distributed. Developers are much more likely to be displaced than networking professionals, for instance, and within development, seniors will feel less of an impact than those just now taking their first steps.
A certain amount of bitterness may be justified, too. After all, it was not so terribly long ago when any young person with ambition in the arts was summarily dismissed by politicians harping on about how they should ‘learn to code’. A better imperative might have been: find something, whatever it is, and get serious about it.
