You are fired

AI has completely exploded since April of last year. What does this mean for the music industry, for example? Will artists soon deliver only their voice? And you've certainly heard of ChatGPT. It's a bit scary today what kind of future this could lead us to. In one study, an AI was able to assess skin tumors better than dermatologists. Will our doctor visits soon be replaced by full body scans from a machine? What will happen to us now? Will AI do away with us all? Will bots take all our jobs and make us useless? How do we make sure things don't go really bad for us?

Is your job in danger?

Yes, your job is in danger. It's not worth pretending it's not. You could now list countless examples of what AI can already do. But let's make it concrete. Vice recently published an exciting article on over-employed hustlers. In the article, various people report how they are doing up to four full-time jobs at the same time thanks to AI help. Challenging, well-paying jobs. ChatGPT generates business plans for you, blogposts or builds Excel spreadsheets.

But if these people are doing 3 to 4 jobs at a time with AI help, doesn't that mean there would actually just be a need for fewer workers in these companies? According to a major study by investment bank Goldman Sachs, about two-thirds of all occupations are affected by AIs. In most of these jobs, bots and co could take over about 25 to 40 percent of the work in the long term. But there is also a small portion of all professions where programs could soon do half or even more of the work. Examples of this are technical support and customer service in general, as well as office workers who primarily enter data somewhere or organize appointments. There will simply be far fewer of these jobs in the long term, according to the study. Almost 1/5 of all global work tasks could be automated. In Europe and the U.S., the figure is closer to a quarter. 7% of all people currently working could lose their jobs in the next ten years due to AIs. And such estimates don't come from a single study. The EU, for example, also assumes that AI will eliminate numerous jobs.

That sounds pretty scary at first, but as is often the case, it's only one side of the coin.

No, we won't all be out of a job

More generally, AIs will have the most impact on jobs that happen at desks. This is simply where artificial intelligences can be deployed most advantageously. Bus drivers don't have to fear for their jobs because a massive fleet of autonomous buses would be insanely expensive. For office jobs, you can use AIs like ChatGPT directly.

Numerous professions need to adapt to AI. Whether lawyers, journalists, or people in administrations, mathematicians, accountants, programmers, writers, and interpreters. But AIs will tend to simply take over the less exciting tasks. Looking something up, pre-formulating something, restructuring things. They are, above all, digital assistants. You can then focus more time and more energy on the actually exciting aspects of the job.

Lawyers, for example, may be able to leave dry routine tasks to the AI in the future and use more time for client meetings instead. Journalists can devote more time to interviews and investigative research. Doctors can get help with prognoses and focus on treatment. The danger that AI will now lead to mass unemployment and then millions of people will sit there and suddenly have nothing to do is very unlikely.

There have always been fears due to change

People have had the, fear for 100 years that eventually all jobs will be done by machines. Automation is scary. In the 1920s, people were afraid cars would put them out of work. In the 40s and 50s, it was already using more and more machines in factories. In the 60s and 70s, photocopiers were declared poison for writers, publishers and academics. In the 80s and 90s, it was computers.

Statistically, few became unemployed, but people actually found new fields of activity within the company. They were retrained, upskilled or repositioned. It's like so often with new technologies, they make old jobs disappear and new jobs appear. It happens all the time, all the time. 60% of all U.S. employees have jobs that didn't exist in 1940. In fact, when new jobs are created, it's almost always because of new technologies.

How is the job market developing at the moment?

AIs are already creating thousands of new jobs every month. The problem is quite simple. It's very easy to imagine a machine destroying an existing job, but it's incredibly difficult to imagine new jobs that don't exist today.

What do critics say?

But there's an opposing theory here, too. This time, they say, automation is different. We would really see the rise of the machines this time. Will everything be automated at some point? The Future of Life Institute has called for a halt to the development of AIs like ChatGPT, with great media impact. Numerous well-known figures from the tech industry have signed on. Among other things, the letter says should we automate away all jobs, including fulfilling ones? Shouldn't we develop human intelligence that could eventually outnumber us, outsmart us, make us obsolete? Replace us?

And there it is again, the fear of a dystopian future. A world in which only a handful of people work and everything else is taken over by machines and artificial intelligences. At some point, we humans will no longer be able to control the AI. In the past, we were always able to save ourselves through innovation and stay relevant, but in the age of AI, we can't do it anymore. There are new jobs, but nowhere near enough to make up for the jobs lost. The incredible potential of KS is too great and computing power is increasing exponentially. There is a danger that the technology will get smarter and smarter.

What do supporters say?

There is also the counter perspective. Mankind will remain in control, AI will lead us to a kind of utopian future. It will solve all our problems.

Such a way of thinking is called longtermism. Longtermists often try to answer existential questions of distant future scenarios today. No one can predict whether we will end up in a dystopia, utopia, or some other scenario 50 years from now.

The debate, even if certain arguments may convince you more than others, is complete speculation. Nevertheless, these highly speculative arguments are sometimes misused. In the here and now, it really has to be about facts. People like to pretend, but programs like ChatGPT are not intelligent beings and have no consciousness. And we should not discuss them as if they were. It's easy to misunderstand the word intelligence here. They are algorithms that can recognize patterns in gigantic data sets and create new patterns from them. In doing so, they do not know or understand what content they are generating. Open letters about development stops and fatalistic lurid visions of the future are just a huge red herring, according to many critics. It's easy to argue about fictional scenarios 50 years from now. The massive problems that AI already brings today can then be ignored. But that would be a huge mistake.

Is AI making our society more unjust?

The models are only as good as the data set they are fed. ChatGPT, for example, was trained with text files from the Internet. And many texts on the Internet are discriminatory, incomplete and badly wrong.

So who do you hold accountable? According to experts, the responsibility would lie with the development teams. You selected the data sets and trained the model, accordingly you have to ensure that your systems are fair. But currently, the way major developers work is still pretty much a black box. The datasets and the way the models are trained are often a secret. Many experts are calling for more transparency to better regulate Open AI and co and hold them accountable. Open AI tries to filter ChatGPT's responses so that the boat doesn't write racist things, for example. But there are quite a few ways to get around these filters. There is also no particularly impressive technology behind this filtering, but simply quite a lot of human work. That's why Open AI and co often rely on low-paid workers in developing countries. They categorize and review content much like Facebook and other social media platforms. Until early last year, for example, a Kenyan team looked at texts and images of disturbing content for Open AI and categorized them. The respective models can then be trained accordingly not to produce such things. The activity was partially compensated with less than two dollars per hour. The cooperation was terminated earlier than planned by the Kenyan side.

According to leading researchers, the current ki industry is based in part on a globalized exploitative system. People in developing countries in particular suffer as a result. Thus, AI products potentially reinforce structural inequalities. AIs could also have such effects within Western societies. AIs, as mentioned earlier, are a threat especially to creative industries, office jobs, and assistant jobs. Making an appointment, briefly looking up a piece of information, creating a table or a presentation, or writing a text. Digital ki assistants will probably soon be able to do all this work. Top earners, however, are not so at risk and may even benefit. The lawyer, the doctor, or the entrepreneur can probably reduce their costs with AIs. Many employees, on the other hand, will potentially have to find new jobs. Thus, it is primarily those who are already well off who could become richer through AIs. The work of the stagnating middle class could be further devalued. This makes it all the more important to provide workers with new opportunities and retraining. AI will not do away with us all. But it will permanently change our working world. It is up to us and our governments to tackle the challenges we already face without getting bogged down in debates about the distant future. The gap between rich and poor is growing globally. AI should not exacerbate this trend.