Artificial intelligence will save the world
An interview with Fabian Westerheide
Fabian Westerheide is a managing director with Asgard - human VC for AI, the AI conference Rise of AI, and a coordinator for artificial intelligence with the German Startups Association. He is an entrepreneur, venture capitalist, and author, and his investments in artificial intelligence mean that we are working less, living healthier, and that we will do better in general. In the past Mr Westerheide was involved in 35 digital companies as an entrepreneur and investor, including Delivery Hero, Point Nine Capital, Team Europe, parlamind, Micropsi Industries, and Accelerated Dynamics.
What actually is artificial intelligence?
“Artificial intelligence” is a top-level term. It means excellently trained digital experts who can perform significantly better than we can. It means collective intelligence, in which all of the members can process the knowledge immediately, for example in swarm technology for drones or in autonomous cars that communicate with each other. It deals with tasks that humans previously had to do themselves: From nursing care for the sick or elderly, stock market trading, or answering e-mails - AI can do it all. The subject is simply extremely large and more than just a technology that you can use to earn money. It impacts society, politics, the military, and of course also the economy.
Can we compare it to the discovery of fire, the invention of the wheel, or printing? Or is it even more?
You quite simply can’t compare it. We started with computers in the 1940s, with AI in the 1950s, and now we are in an acceleration curve, building AI systems which are superior to humans in certain specialist areas. We are creating intelligent life forms in a variety of versions.
This could possibly be compared with human development two and a half million years ago. But that has required a large number of iterations. It took a while until homo sapiens emerged 60,000 years ago. However, this long period resulted in the creation of what makes us special: our emotional and cognitive intelligence, our spatial reasoning, our eyes and our sense of smell, our logical reasoning, our imagination, our language.
Every single one of us has a wide variety of forms of intelligence and has many abilities that all work together. We are trying to reproduce this in a single generation. We can’t do it without the Internet. I believe that artificial intelligence is the logical continuation of the Internet. The Internet has given us access and has also generated the infrastructure and the requisite data. The Internet will continue to grow together even more closely with the aid of artificial intelligence. It will ensure that we spend less time at the PC and accessing the network, and machines will increasingly approach us instead. As we have already seen with the smartphone: This device is not just a mini-computer, it gives us access to a host of opportunities - right there in our pockets.
Why do you invest so much in artificial intelligence?
Artificial intelligence will save the world. I would even go so far as to say that it is the second cognitive evolution. We are creating a whole new form of intelligence. We already have artificial intelligences that are much more intelligent than humans and can beat us at playing chess or Go.
Our imagination is limited - so is it even possible to imagine what could possibly happen?
Partially, yes. Let’s start with what we can imagine.
- For example farmers use artificial intelligence. They use a networked sensor which communicates via the Cloud and satellites to measure the ground’s phosphate and oxygen content. If the farmer has a large farm, this can make a lot of sense, because he no longer has to fertilise the whole field and only has to spray pesticides where he really has a problem. You could also do that via satellite monitoring. The AI system evaluates the photos and says, as a preventative measure, what has to be done in order to optimise the output. That is an intelligent service. Otherwise people would have to actually go there and take a look.
- Thanks to artificial swarm intelligence self-driving cars can communicate with each other and regulate their speeds so that there are no longer any traffic jams. Or they can tell each other where they shouldn’t go because there’s a traffic jam there.
- We can also imagine AI systems that play chess and are excellent at solving logical tasks.
What we can’t imagine: a so called AGI, an Artificial General Intelligence, or the same thing that humans regard as being on a par with themselves. The moment in which you have a machine which works intelligently at a human level, that can speak, communicate, hear, see, and learn like a human does. It will probably download all of the knowledge which is digitally available the very next day and learn from this. And on day three this intelligence will probably be more intelligent that the entire human race put together. That could happen. And when it does we can’t understand it any more. There are a couple of people who are dealing with this such as Stephen Hawking or Bill Gates. That could happen in five years’ time, but it could also be the case that we haven’t even managed it in 20 years. However the intelligence which we don’t understand is coming.
It’s already the case that you can’t tell a self-driving car what the command is. You can’t tell it how it should learn. You can’t see how these machines think. You can only remove the so-called black box with a great deal of effort and then attempt to decode the neuronal network. The result is hundreds of pages of code. Then you could read this code and attempt to understand it. But right now it’s like in a brain: We don’t know how the decision was made. However that is also a major plus side - these systems can learn. You don’t have to tell them: These data provide this result. Instead you tell them: Here are the data, we’d like this result, find it out yourself.
Could horror scenarios like in the Terminator films happen - perhaps these aren’t even too far off?
I am not a fan of scenarios like the Terminator films - because they scare people. However AI shouldn’t make people afraid - artificial intelligence will mean and is already ensuring that we are working less, living healthier, and that we will do better in general. The smartphone is the best example of this if we use it properly. It helps us to save time in communicating with others - something that we used to have to invest a lot of time in. Writing a letter used to be very time consuming - if you spelled one word wrong, you had to re-write the entire page so that it didn’t contain an error. Then the letter had to be transported - which also cost a lot of time. Now you can dictate the same message into your phone and have it sent, and that takes just a few seconds. There’s an awful lot of intelligence involved in that.
However the future will be more like humans working together with machines, not machines alone. Doctors use artificial intelligence to scan X-rays for cancer. Even today they have 20 years of experience with specialist training and still have a five-percent error rate. However AI can diagnose this with an accuracy of 99.9 percent. Results are available faster, at a lower price, and in better quality. We will live longer, because we will receive better healthcare advice with medical support from AI systems. However the consultation is always performed by the physician. The doctor will continue to always take blood, interpret the results, advise patients, and issue a recommendation based on his experience. That will be the case in many other areas too: It will always be human plus AI.
We’re not going to consider Skynet from Terminator, because that is a military application. We will always have to respect the military, because the military will never focus on the wellbeing of humankind. Of course AI has the potential to be a weapon. But we must never forget: AI systems learn from us. You shape them with the data that you input. It’s a bit like with children. You can bring a child up to be a good person, and you can bring a child up to be a bad person. A child will however never become a bad person, because he was born that way. It’s exactly the same with AI. It’s up to us. We can build good or bad AIs. They start out as a blank page.
On the other hand, of course, it is also a question of motivation. For now AIs are not greedy, corrupt, or power-seeking. That means that the AIs don’t have these bad or good human characteristics. In addition they don’t have any intentions of destroying or damaging human kind.
There will be cases where we will have to work against something. For example if the financial industry gets out of control as a result of an AI, if it sells all kinds of junk mortgages thus brining the whole system down, then we will have to counteract. These systems are all made by people, the mistakes are thus made by people, and the AI problems which also occur will be human problems. Because humans have given them incorrect incentives and haven’t programmed them correctly. We will learn from this and help to control everything. Of course the regulations and supervisory authorities will have to play their part - this all has to be in place in good time.
Why do you think that Deutsche Telekom is the right company for AI?
Deutsche Telekom is a telecommunications company. It has contacts to B2B and an excellent strategic position to become a leader in this sector. If Telekom can’t do it, then who else can?