When will AI make me an unemployed accountant?

In reality, they're stupid tin cans

It is said that artificial intelligence will fundamentally change our world - just as electricity once turned everything upside down.

Yes, AI will change our world and our system. It means the end of the industrial age. In the industrial age we are dependent on finding a desired solution. Then we try to standardize them together, do marketing so that people get roughly what they want - or that they want to have what they think they need. Then we build factories and processes very expensive because we know we can sell a lot of that stuff. With AI, we no longer need this economies of scale. We tell the machine what we want and it always does it in batch size 1, i.e. also in one-off production, always at an affordable price. This means that the economies of scale will no longer be relevant. And that completely changes the economic rules of the game.

AI is turning our economic system upside down - and how will AI itself develop over the next few decades?

It is very difficult to say. What can be said, however, is that if we apply today's techniques, we can replace a large part of our procedural activities. That has a huge leverage.

Do you have an example of this?

The efficiency of autonomous driving. That probably brings us fifteen times the usefulness of vehicles. It's greener, faster and you can sleep and work while driving. Considering that a third of the economy consists of logistics, that is enormous.

The development is happening so rapidly at the moment that many people are afraid that we will soon be controlled by an omnipresent superintelligence. Professors from renowned US universities and even visionaries like Elon Musk warn of the danger of this control.

I find these worries completely absurd. I wouldn't say it will never happen, but anything that goes in that direction is definitely 150 years away. Therefore, one should rather take care of coping with the changing economic system. Of course, machines can outperform us, and they already do that in many areas. That has always been the goal of machines. Nobody wants to be a crane by profession these days, they lift everything better and more easily than us humans. Basically, it depends on us. We could have blown ourselves off the planet with the atom bomb years ago. You could do that very efficiently with AI, but you don't have to. So far we have also managed to prevent our own destruction.

The threat may be located where the raw material of artificial intelligence is: platforms such as Facebook, Google, Amazon, Alibaba and Co. have access to huge amounts of data. Will monopoly formation be a big problem?

We humans invented these platforms because we wanted to. None of the authors of the dystopias "1984", "Clockwork Orange" and "Brave New World" could have imagined that we would implement and use all these techniques without a dictator. We did this for comfort - and now we are like the sorcerer's apprentice who doesn't know how to get rid of the ghosts he called. We should deal less with the platforms and more with the things that the platforms do not yet occupy. Building another search engine is difficult.

Platforms like Amazon have become an integral part of everyday consumer life. The virtual purchase assistants that we know seem to act in the interests of the customer, but the company's sales interests are behind them. Does it have to be like that?

We always assume the assistants are acting in the interests of companies. At the moment we manipulate in the interests of companies, earn a pig's money with it and call that advertising or influencer marketing. In my opinion, advertising is much more manipulated and sometimes even ripped off. In the assistants' business model, you don't earn money with someone's transaction, but with your life. The worst for a provider would therefore be if a customer switched to another provider. I believe that the assistance systems make people much less susceptible to marketing.

Nevertheless - the fear of surveillance is great, fueled by the many reports from China, keyword social scoring. Do we have to expect this form of control or surveillance in Germany?

Of course, the technology can be used for this. I have long been concerned about what data is being collected in the name of counter-terrorism, even though the likelihood of dying from an act of terrorism is very small. However, this explains why all of our data is recorded and archived. AI helps make this archive easier to search through. I don't think it really does anything for our security. Models of total surveillance will definitely be possible. The problem is not in the technology, but in humanity. The interesting thing about China is that the majority of people find this social scoring good because of the completely different social system. I hope that we in Europe will remain a society in which we find this kind of thing to be the very last thing. That would be a way to attract young, hardworking and thinking people.

Should we therefore restrict research on AI through laws?

You can do good and bad with AI. So it is with everything. Often the response is to regulate research. As a result, however, the researchers simply emigrate and continue elsewhere. But the only way to enforce your values ​​is to be at the forefront of research. It is much smarter to be two steps ahead of the others. So I find the approach of downgrading research completely pointless, because that is a guarantee for a catastrophe. Because if the economic situation shifts, the state no longer has the say.

Let's talk about the human being who is being degraded in the whole discussion about artificial intelligence. What can humans do better than any machine?

Humans are the only ones who have a will. People not only replicate reality and the past, but actually say they want to change something. People have an artistic urge and swim against the current. They work things together that don't really belong together. People also take risks. We call all of this creativity. This is exactly what makes people seriously happy. When you create something, you can't compare it to a like on an Instagram post. I don't know if it is even possible to ever create a machine with this will of its own. If so, it will take a long time.

So wouldn't you describe Alpha Go's moves against the world's best Go player as creative?

No. A machine can also draw a Rembrandt. But that doesn't mean you can introduce Cubism. In the game of Go, many have called the machine's steps creative because they were surprising. However, the machine has simply learned different and more patterns than a human can remember. It's not creative, it's consistent. So does the machine only outperform us in dull, monotonous things? You could say it like this. Machines can evaluate X-ray images consistently and do it with the same quality seven days a week, 24 hours a day. Machines are responsible for all the repetitive stuff we humans weren't made for. That's the great thing about them.

Then we should actually be happy to have machines. What needs to be done so that machines with AI are accepted in Germany?

That's what I worry about every day. It is completely unclear to me why they are not accepted. But I think it's because of how we explained these machines. By declaring them to be intelligent and suggesting that they work like brains, which in reality they don't, we have created an image of fear in people. In reality, it is still stupid tin cans that relieve us of work that does not make us happy anyway. When I see people trudging into their offices in the morning, I rarely see happy people.

How many of these monotonous processes can AI take over in companies?

You are now asking the believer. In my opinion, you should actually automate everything that is not interpersonal. Around 80 percent are possible nowadays. But I say that now as a technology disciple. According to the World Economics Forum, 29 percent of the work that humans used to do is now done by machines. The forecast for 2025 is 53 percent.

Don't many people then become unemployed?

It is always said. But I think that's 100 percent illogical, because if we all became unemployed after the machines took over our current work, that would mean that we are already doing everything that is necessary today. That can not be. We have a huge innovation backlog, the planet is collapsing. Our society has succumbed to egoism to such an extent that we can now choose extremism again. There is an upside-down age pyramid, a broken pension system. There are so many important things in the world that need to be fixed and made better. So I think this fear of mass unemployment is idiotic.

That sounds utopian. In virtually all revolutions there was a lost generation who had a hard time before the renewal bore fruit. Is that not going to happen with AI?

That's true. Usually the transition with something like this is very painful. Otherwise it was always the case that the innovation came, which is why jobs were then cut. The money earned was saved and only invested again over a long period of time. Now, however, the platforms have such a huge lead that you have to reinvest directly in order not to lose touch. This creates new jobs and the transition could happen without a global economic crisis or something similar. This is an unprecedented opportunity.

What could a new job look like for a 45 year old accountant who has done nothing else in his life and loses his original job?

I wonder why they didn't use the taxi or truck driver as an example. So, first of all, the question arises whether he enjoyed doing this job. Second, he could train the machines, for example. Certainly jobs and entire industries will no longer exist at some point. Hard coal mining has only now been abolished, although 40 years ago it was clear that it would be completely uneconomical to promote it through subsidies. Even people who had no prospects were trained. These industries should not be kept alive convulsively, but should only be financed for a generation to retirement. As an occupation, the last of their subject could accumulate knowledge and feed it into the machines. Who knows when and where you can still use these skills.

When jobs are lost, new ones are needed. What jobs will there be in the future?

Nobody can predict specific job profiles. 20 years ago nobody could have imagined that Youtuber could one day become a profession. Basically, the professions of the future must be thought more humanistically. It's about what people are made of and how they can use it to advance society.