Artificial Intelligence | Less work through AI – actually a good thing

New technologies and the preservation of jobs are something like natural enemies, probably since the invention of the printing press at the latest. Anything that makes work processes more efficient has the effect that people have to work less. This practically happens automatically. In capitalism, the one side then aims to make everything cost less. After all, they think, why should they pay people for something they no longer do themselves? While the other side has no interest in being replaced by machines as long as it doesn't mean a reduction in working hours with full compensation. The result: poverty, misery, labor disputes. Two steps back, half a step forward.

We're currently experiencing a new phase of this dynamic with so-called artificial intelligence (AI). So-called because there is no such thing as artificial intelligence—that can't be emphasized enough. It's a collective term for many different types of software, which would take too long to explain here. But none of this has anything to do with human intelligence.

It's no surprise that AI, like digitalization and automation, has the potential to endanger jobs. At the same time, many new jobs are being created: people who deliver ordered food and many other things; people who write online reviews; people who watch and filter violent videos before we see them; and, more recently, those who prepare data for things like ChatGPT. Barely protected, incredibly poorly paid: (More) labor struggles, attention, and solidarity are urgently needed here.

» Competitive Sports« AI Dystopia

The topic isn't yet so present in Germany, but things are different in the US. Connie Loizos, editor-in-chief of TechCrunch, a US tech news website, called predictions about AI jobs the new "competitive sport" this week . Representatives of major tech companies are quick to make announcements about it. Ford CEO Jim Farley believes that half of all office jobs will disappear. The language learning platform Duolingo, which is also popular here , announced in April that it would no longer place contracts for things that can be done by AI. (That didn't go over well .) And Microsoft has just announced that it plans to lay off four percent of its workforce .

What's new is that people are now talking about those who develop the software themselves. Generative AI (ChatGPT & Co.) can not only create university assignments and application letters, but also write software. Is the digital revolution now consuming its own children?

Now, software development is more than just translating commands or database structures into programming languages: Here, too, experience, an understanding of complex interrelationships, knowledge of the chaotic IT patchwork in companies and administrations that "just happened that way," and much more are important. It's likely that simpler tasks, more commonly performed by entry-level professionals, will be taken over by AI applications, but more complex tasks are not so easily replaceable. In addition, AI tends to be unintelligent, constantly makes mistakes, and therefore always needs to be checked, which can be time-consuming. A new research paper from the University of Oxford concludes that AI writes significantly worse code than previously thought.

Just a year and a half ago, the traffic light coalition government stated in response to a question from the CDU/CSU that it "continues not to assume that the use of AI will have a significant impact on the unemployment rate." It also stated: "Strengthening the rights of works councils and staff councils in the use of AI systems is of great importance." It added: "In addition, in the negotiations on the AI ​​Regulation, the Federal Government is advocating for a clarifying opening clause in the AI ​​Regulation, which would allow for more favorable member state regulations with regard to the protection of employees when using AI systems."

It is not AI that is the problem, but deregulation

Whether AI will actually lead to job losses or whether others will be created instead, and for whom: some studies say one thing, others another. However, there is clearly a discernible interest among tech companies in fueling this debate. It's possible that this is just the same old flash in the pan: we're seeing ever-new tech buzzwords in ever-shorter cycles that supposedly turn everything upside down. Does anyone remember NFTs? Bitcoin? Cryptocurrency? Blockchain? US tech journalist Brian Merchant received so many responses to his public question about who has experienced their job being replaced by AI that it has evolved into a comprehensive project. So, none of this can be completely dismissed.

The new federal government—with the same SPD in the same Labor Ministry that was responsible in the previous legislative period—is now singing the tune of tech companies, who have one primary interest: They want less regulation and less bureaucracy. The new Digital Minister, Karsten Wildberger (CDU), just made this very clear in a speech given by the IT industry association Bitkom: His goal is "more practical implementation of regulation." Translated, this means: Companies should be less restricted by regulation. And reducing bureaucracy simply means less documentation, (even) less transparency in administration, and less protection of workers' rights, fundamental rights, and human rights.

US President Donald Trump was one of them: In the recently passed "Big Beautiful Bill," he wanted to ban any regulation of AI for ten years – a move that the US Senate narrowly prevented. This would have affected not only jobs, but also the use of AI. To name just one example among many: automated decision-making in job application processes, which cannot be challenged because it is completely incomprehensible why the software made a certain decision and not another.

Of course, tech entrepreneurs dream of replacing humans with software with as little regulation as possible, without having to worry about supply chains or workers' rights. They're happy to take a few subsidies, reduced electricity prices, and incentive programs for the next buzzword. The products and services they want to sell then probably won't be as good as before. Wouldn't it be nice to actually be able to talk to someone there after a misdelivered package from Amazon or DHL to clarify the error? Or to pay only half the price when the service is so much worse than it used to be?

At the same time, there would be nothing wrong with leaving some of the work to machines – as long as it's regulated so that everyone benefits: less work, more time, but at the same time, enough income. But that won't happen without new labor struggles. And a few monopolies would probably have to be broken up first.