This is how you recognize real AI experts

Market researchers are familiar with Haßloch, the “largest village in Germany” in the southern Palatinate with just 20,000 inhabitants. As a place that perfectly represented almost all demographic characteristics of the entire population of the Federal Republic, it served the Society for Consumer Research (GFK) for many decades as a test terrain for investigating consumer behavior. That came to an end in 2016, because GFK also had to realize that complex relationships simply cannot be fathomed in this comparatively banal way.

The situation with the OMR is very similar these days. Nobody could get past the Hamburg Westermeyer Fanboy and Girl Festival last week. Whether you wanted it or not. You didn’t even have to be there to do it. All social media channels were clogged with impressions from participants or comments from non-participants.

The unmistakable cacophony on the subject of AI, which found its provisional live highlight at the OMR Festival, sometimes gave the impression that 70,000 AI experts were setting the course for the weal and woe of society and the economy. Seen in this way, the OMR was representative of the whole republic, which after 80 million national trainers and virologists now seems to consist of 80 million AI experts.

Everyone has something clever to add to the discussion, and yet in the end it only serves to speed up the general gasping. Although this seems to have its origin in roughly equal parts in existential fear or gold rush.

It would be so important right now to take a closer look. to argue in a more differentiated way. But above all: to know your way around. Or to use the quip of an otherwise less quotable “cabaret artist”: “If you have no idea, just shut up.”

In fact, I firmly believe that in the AI ​​debate, we are surrounded by a mass of vagrants trying to explain to those who have been studying the subject intensively for many years how to do their job. What is already difficult to endure on a daily basis with Linkedin and Co becomes unbearable at specialist events and even more so on the political stage.

I would therefore like to give a few tips on how to separate the wheat from the chaff and how to identify real experts. Above all, there is the equation of artificial intelligence with automation. This correspondence is simply not correct! Because in the vast majority of cases we are talking about AI applications that take on individual activities – this also applies to the very strong focus on generative AI à la ChatGPT or Midjourney. Only subtasks are replaced here, but not entire processes.

That means: without people it doesn’t work. Human beings are initiators, process monitors and/or finishers – often all three. No result without prompt, quality control and/or human defined use of the result.

Even automated repetitions do not exist without a corresponding impulse to optimize or identical repetition of the corresponding activity. Sorting out black, rotten or unripe French fries using AI-based image recognition and then automatically packing the flawless potato sticks in bags is a lot further. Here, no one is needed in the entire process.

The original hope that AI would make our lives easier or automate and accelerate entire processes has not really been fulfilled so far. There is no real relief through independent automation. Neither quality assurance nor personal initiative work. Generative AI (aka “the machine”) doesn’t handle anything on its own. Humans are still in the loop. This raises the question of which evolutionary step we are currently in?

In any case, we are still a long way from automating processes through artificial intelligence – especially in the creative field. However, the current debate leaves most people with a completely different impression. Without making a crystal clear differentiation, however, it is difficult (at least for me) to have a real discussion about the extent to which AI is actually finding its way into our lives and work. The fact that many participants in the fundamental debate about AI are usually unable to distinguish between machine and deep learning and stumble at the latest when it comes to neural networks is a huge annoyance for me personally.

Whether politicians, journalists, keynote speakers or panellists – please do your homework if you want to take part in the debate! Otherwise this debate will degenerate into a superficial argument, as we are all too familiar with from the corona pandemic or the retirement of the national soccer team. And then it’s really better to avoid events or new hate holes like the OMR Festival in the future.