"Innovation needs people"

Chris Boos, one of the most sought-after German experts on artificial intelligence (AI), claims that innovation will always need people. In this interview he explains why the disappearance of many types of jobs presents a huge opportunity—and could help solve the really important issues.

Interview Thomas Schmelzer  Photos Lottermann und Fuentes

Mr. Boos, your business is automation with the help of artificial intelligence. How do you reassure your employees that they’ll still have a job five years down the road?

We don’t have to do that at all. Particularly those companies that deal with artificial intelligence are constantly asking themselves how to make things as pleasant as possible for their employees. We are dependent on our employees because we know exactly what the limits of artificial intelligence are.


Artificial intelligence completely fails when it comes to developing new ideas. That’s something only we humans can do. Unfortunately not on demand, either. You can’t tell your employees: Monday, 8:30 a.m., please deliver an idea! We can only create an environment where people feel comfortable—and then hope that they’ll do what they remain unbeatable at.

What exactly would that be?

I divide it into two large fields: creativity and social interaction. Being creative means that you sometimes swim against the current, that you keep tinkering until something works, and that you take risks so that you can claim the glory as a pioneer when you succeed. No AI in the world can do that. Without people, there’s just no innovation. In social terms, machines do even worse. Nobody wants to be waited on by an emotionless robot in a restaurant.

Quite a few researchers have been working on fairly empathic machines for a while now.

Of course I can program a bit of friendliness into a machine, but that’s just a simulation. Machines are nothing like actual people. That’s something anyone who has ever had to call an automated hotline notices. Machines have different strengths.

Better at linear, repetitive tasks, they say.

I think that’s true, by and large. In very simplified terms, I’d say everything that’s difficult for people is easy for machines. And vice versa.

Which results in the famous thesis that artificial intelligence will keep the boring tasks at bay in the future—so that we’ll finally be able to take on the exciting and creative things in our work.

That’s the great hope, and I think we’ll be able to fulfill that for the most part. Take IT for instance: there are many complicated processes and that’s why administrators are inundated with user queries. Their actual job should be to keep the systems as secure and stable as possible. This for instance is where our AI comes into play, taking on routine tasks and freeing up time for administrators to do their real jobs.

With the result that there will soon only be five administrators in a company instead of twenty.

This is exactly what has almost never happened for us. Only one of our clients cut jobs. All the others improved their service or gained and served more customers with the same team. Ask an IT admin how much fun it is to spend hours resetting passwords every day. That happens all the time on a daily basis. And each system has a different procedure for it. This sort of task certainly doesn’t need a very well trained and consequently expensive IT expert. AI can already do this better, faster and cheaper, making it more efficient overall. And the administrator is finally liberated from this tiresome task.

Yet people are still fearful about their jobs.

Let’s not beat around the bush: many jobs will die out in the future. I would say up to 80 percent. Yet at the same time new jobs will be created—it’s an enormous opportunity.

It’s news like this that makes even more people worry.

Because they don’t see the opportunities. Banking, for instance, is an industry that will be swept away by disruption in the coming years. An unbelievable number of people are busy handling internal processes. Deadly boring, always similar tasks—and therefore ideally suited for AI. Machines will be taking over more and more processes here. If the banks are clever, they’ll use the resources that have been freed up for customer service—something they’ve been neglecting for many, many years.

What about those who really enjoy in-house controlling?

They make the best experts to manage the transition to automation. They know the processes blindfolded and can therefore perfectly adjust, train and continue developing the AI. Instead of just doing monotonous tasks, they’ll finally become creative and can use their knowledge to improve things. By the way, this isn’t some new process. When films with soundtracks replaced silent films, a number of organists, who had previously played live music to accompany the films in cinemas, were suddenly out of work. They then invented film music—and became pioneers in this new field.

„Everything that’s difficult for people is easy for machines.“

Yet you need far fewer people producing film music than you need organists in cinemas.

That’s true, and it doesn’t get us anywhere to claim that everything only gets better with machines. We’d be kidding ourselves. But there’s no question that we will always need people. I think that many people are scared that they won’t be needed any more or will have to start all over again from scratch, but it won’t be like that at all.

What will happen with today’s cinema organists who don’t want to reinvent themselves as film music producers?

They’ll most likely find jobs in customer service. The more decisions made by machines, the more important service becomes.

You’ve recently been appointed to the German Federal Government’s Digital Council. What does Angela Merkel think about your tips?

Well, we’re not a program council that develops an agenda and then confronts the Federal Government with a list of demands. But I do notice how much everyone is listening and is interested in our opinions.

There’s a long road from listening to taking action. What are some of the first things governments should start doing in the coming years?

We finally need real change. Our society and our economy are still geared far too much for optimization. So far we’ve continued to perfect the mechanisms of industrialization. Dividing processes into small steps, exploiting economies of scale, developing efficiencies: we’ve become true masters of these. What’s missing is genuine innovation.

Aren’t the intelligent machines we’ve been talking about the whole time an innovation?

Yes, but much of what is sold as AI these days is unfortunately just pure marketing, nothing more than old wine in new bottles. Sometimes it’s just an optimization, because artificial intelligence is simply unbeatable in optimizing. Systemic change this is not. The more I deal with artificial intelligence, the more it becomes clear to me how little we actually know and are able to do so far. We know as much about a true human replica as people from the Stone Age knew about how smartphones work.

Is the internet one of these pseudo innovations?

Yes, basically. There were communication networks a thousand years ago. We’ve ultimately done nothing more than accelerate them. First there were horseback couriers, then telegraphs, then telephones and now we’ve got the internet and Facebook.

Quite a lot of people think that Facebook is a big revolution in communication.

Facebook is rumors. We know about those from antiquity. Now they’re just being spread more rapidly. A lot has changed, of course, but the systems have remained the same. It’s no comparison to the radical change from manual labor to the steam engine.

What would you consider a true breakthrough?

Nuclear fusion would bring enormous progress.

Will AI help us finally move forward?

Only indirectly. Artificial intelligence can help clear people’s minds and give them the time to deal with the really important issues of our era. For instance with creating clean energy, or finding solutions to climate change.

So we humans must continue to do all the work?

I’m afraid so. If that’s not good news, I don’t know what is.