– Wherever technological development has occurred, Cassandra's voices have been heard, predicting doom. However, mass technological unemployment has never occurred – says Jacek Mańko from the Leon Koźmiński Academy
Three-quarters of knowledge workers already use artificial intelligence, and more than half of them fear that using AI to help with important tasks will make them no longer indispensable, according to a study by Microsoft and LinkedIn.
There is no shortage of predictions about the impact of AI on the job market, and most of them are written in an alarmist tone. Experts warn that AI will put many people out of work. According to a report by Goldman Sachs, up to two-thirds of professions are to be replaced, which is the equivalent of 300 million full-time jobs. Is there anything to be afraid of?
Layoffs or greater productivity?
– There is a lot of pressure for automation – says Jacek Mańko from the Department of Management in the Network Society at the Leon Koźmiński Academy. He adds, however, that the assumptions of other companies researching the subject are sometimes more optimistic. – For example, the consulting company EY predicts that generative AI can increase productivity, bringing the US economy profits from $650 billion to as much as $1.2 trillion within a decade, which will translate into additional growth GDP by 2.5–5 percent – indicates Mańko.
According to the expert, the predictions – both those predicting that most jobs will be taken over by AI and those indicating an increase in productivity – are quite extreme and rather exaggerated.
– Wherever technological development has occurred, Cassandra's voices have been heard, predicting doom. However, mass technological unemployment has never happened – he emphasizes. – Almost a hundred years ago, economist John Keynes predicted that technological development would lead to us working 15 hours a week in 2030. This has not come true, on the contrary, there is the impression that we are working more and more.
Or maybe this is where new technologies will help us – they will relieve employees, take away boring, repetitive duties, and people will be able to work more effectively? Mańko doesn't have particularly good news here either:
– The renowned American economist Daron Acemoğlu argues that technological innovations do not always contribute to increased productivity. He cites the example of self-service checkouts – the time that customers spend in a store is the same. On the one hand, cashiers and cashiers hear that their work can be automated and reduced, but on the other hand, in addition to operating traditional checkouts, they also have to assist with constantly malfunctioning self-service checkouts – comments Mańko. – If we look only from the perspective of automation, employees seem to be only a source of costs that can be reduced. However, there are also arguments against such a rush.
The Middle Income Trap
According to the expert, even without the AI boom, there would be several reasons to worry in the labor market. Among them, he mentions waves of layoffs in the IT sector.
– But problems can also be seen in other industries – comments Mańko. – Every few weeks a production plant in Poland closes, others are laid off in groups. This year alone, there have been several thousand liquidated positions. Wages are indeed growing, although still too slowly. This is of course a positive phenomenon from the perspective of employees, but foreign investors are starting to look for other emerging markets, especially those with cheaper energy. We are no longer on the periphery, but we are not an innovative economy in the center either, we are at risk of falling into the middle income trap – he explains.
The year 2024 may be the time to assess the balance of profits and losses. – The situation should be considered in a broad context, as it is influenced by the falling and inflationand still high interest rates – adds Mańko.
The development of artificial intelligence may also affect this situation on the job market. For example, in IT, some tasks can be automated to some extent. – Perhaps junior positions, the so-called “entry level”, i.e. for beginners, will slowly become profitable to transfer to AI – says Mańko.
The decision was made by artificial intelligence
So let's answer directly: will AI replace us or not? – I would clarify this question. Saying that AI “will come”, “take away jobs”, “replace us” is a linguistic simplification – emphasizes Mańko. – Artificial intelligence is not an element, a natural phenomenon, a force that makes decisions on its own. Its development depends on people. And it is people – and in fact the managerial class or, at some level, the political class – who will make decisions (or create conditions) regarding whether, for example, translations or journalistic texts written by AI are already good enough to be able to transfer these tasks to such systems.
Jacek Mańko emphasizes that in the long run these models are not as good as professionals: – Creative circles, for example, are aware of this, but will managers of large companies resist the temptation to cut costs? It's hard to say. Although we would probably still prefer to read content written by people – he says.
But it's not just a matter of preference. In the case of texts or translations proposed by AI, for example, the responsibility for them will still fall on the human. In court, the justification that “the decision was made by artificial intelligence” is unlikely to hold up.
– Technology is currently unreliable at the least desired moments, but we'll see what the future brings – says Mańko. What should employees do in this situation? – Even experts disagree on what exactly awaits us. Are we experiencing a revolution that will change everything? If so, we should develop digital competences, learn how to use AI at work – he lists.
– But these forecasts are so unclear, sometimes even mutually contradictory, and technology is developing so quickly that it is possible to overshoot – he adds. And he gives an example: – Two years ago, it was said that the profession of a creator of queries, prompts for artificial intelligence (a query or command passed to the AI model – editor's note) would be sought after – this requires a fairly precise language that meets certain parameters. In the meantime, tools have appeared that change everyday language to such queries. Training turned out to be unnecessary.
Main image source: Shutterstock/ Willyam Bradberry