Artificial intelligence (AI) and robotics will help create 58 million new jobs, according to a recent World Economic Forum forecast. Scientists present innovative AI developments almost every week. Our colleagues at the business webzine Invest Foresight talked to Jin Kolesnikov, a futurologist, foresight practitioner and the founder of SingularityU Moscow, who shared his views of when robots would become smarter than humans and whether artificial intelligence might be a threat. This interview is published by courtesy of Invest Foresight.
How much time is needed for a strong, super-smart AI system to be built?
First, we need to agree what we mean by “strong artificial intelligence.” In the classic interpretation, we are talking about a self-aware AI. This kind of system is not born overnight – it takes time. But if we mean robots that can pass the Turing test, according to the latest forecasts, they will appear in 2030 – smart systems ranging from smart homes to smart cars. People will be able to talk to them without knowing if they are talking to a robot or a person.
Once a strong AI appears, ethical questions will arise. Will humans be able to keep artificial intelligence under control?
Artificial intelligence works according to the protocols and algorithms written by humans. These algorithms are protected. They are aimed at making robots useful. Therefore, there is no threat from modern robots. If we talk about self-aware artificial intelligence, I think we just will not be able to understand it. Try talking to ants. You will first feel superior to them. And then you will realize that you cannot make a deal with them. Ants are something beyond human control, something completely different. Strong artificial intelligence will solve human problems in the same way, but people will not need to find common ground with it. It is important that AI simply perform its work.
How will humans interact with artificial intelligence?
I think there are three major areas for such interaction. The first one involves unmanned vehicles. Man can help them navigate the urban environment, and be partially involved in the control process. The second area is more complicated; it includes education, where artificial intelligence will give advice to people on better ways to teach others. And the third area involves artificial helper algorithms, such as the IBW Watson supercomputer that understands human language and can solve a wide array of tasks. Such AI can actually replace man for a certain scope of work.
Do you think AI will pose a threat or bring new opportunities in the future?
Any automation provides new opportunities for people. There is no threat. Even such important technologies as nuclear energy have not exterminated mankind. Even with all their irrational behavior, humans are intelligent enough not to annihilate themselves as a species. And in the future, humans are highly unlikely to create something that can kill them.
Can you name any developments that will have the most drastic effect on AI advancement?
Neural networks have vast prospects, but they have already reached their peak. This technology enables to solve various tasks, but there is a limit to its potential. Another promising technology is mixed or hybrid reality, where artificial intelligence works together with man in a single space and there are tangencies between real and virtual objects.