AI will not destroy humanity but right now is like a car without brakes. This is how Sweden’s representative in the newly started UN body for AI issues, Virginia Dignum, describes the new technology.
– It is the human who controls the AI that can become dangerous, she says.
AI has been used to produce the latest, and probably last, Beatles song. AI can also control drones in war.
Recently, Virginia Dignum, as the sole representative of Sweden, was elected to the UN’s advisory body for AI issues; United Nations High-Level Advisory Body on Artificial Intelligence. Among 2,000 experts from various fields and countries, Virginia was selected by the Swedish government.
“Like a car without brakes”
The UN body and the other 37 members will work to find out how to control AI so that the technology is not used in the wrong way.
– AI right now is like a car without brakes, with a driver without a driver’s license on a road without road signs. We have to make sure we find ways to control this car, says Virginia Dignum, professor of computer science and expert.
There is a big risk that AI spreads incorrect information. But AI is also conservative and perpetuates prejudices, according to Virginia Dignum. She lets an AI application produce images of nurses: all the images show women. Pictures of doctors show only men.
– AI does what we do and if most people are prejudiced, so is AI.
But AI is also a tool that can help us make better decisions. At the Department of Computer Science in Umeå, for example, a program has been developed with the help of AI that shows how a pandemic affects people depending on various restrictions.
Concerned about man
Nevertheless, voices have been raised that AI development could become dangerous for humanity and therefore must be paused. Virginia Dignum doesn’t think it’s possible or desirable, but she believes more research is needed to find out how the systems can best be used for a good purpose. She’s not worried about AI taking over the world.
– It is not AI that will want to destroy us humans, I am concerned about the people who will use AI. If a malicious actor can do what they want with AI, it can be dangerous, so we need to find ways to regulate its use, says Virginia Dignum.