Technological advancements are a part of our everyday life. The result of this is more advanced artificial intelligence which could destroy our mankind.
While the evolution of AI to a human-level of intelligence could be extremely beneficial to civilization, if it were not to cater to our needs then the consequences could be dire.
Straight from her presentation at the Brain Bar Budapest where she spoke about the existential risk that artificial intelligence might wipe-out humanity Viktoria Krakovna told Express.co.uk that we need to be careful when creating AI that serves us as it may not understand what we’re trying to achieve.
The co-founder of the Life Institute said: “The risk that human-level intelligent AI is not aligned with human values would be quite a large risk.
“If we built AI that is more intelligent than we are, that can solve the problems that we haven’t been able to solve, like eradicate disease, poverty and war, then it could build a much better world than we have, than we have been able to build with our level of intelligence.”
However, she added that first we need to understand what humans want, but conveying that desire to the computer is probably the most difficult part.
Ms Krakovna said: “If you asked [AI] to get rid of cancer, we think of that as a beneficial goal, but the most straightforward way of eradicating cancer would be getting rid of all humans – if there are no humans, there’s no cancer.
“In order for super-intelligent AI to help us get rid of cancer or any other disease, we would have to specify that we have to preserve human flourishing while getting rid of human suffering so we would have to extract philosophical concepts that are really not that easy to specify to a machine.
“If we end up with a misspecified idea of human values then that could be really bad.”
Ms Krakovna added that rather than AI engaging in all-out war with humans, the real danger lies in it not complying to our needs and the machines ultimately using up all of our resources which could put an end to humanity.
She said: “I don’t see us being replaced by AI, I have more of a concern that we would just be in the way of what it is trying to achieve and it could use our resources for something else. I think we need to worry about value misalignment between us and them.
“I think that’s certainly possible, even if the agenda is originally programmed by us. It is not clear whether super-intelligent AI would use us for something or whether we would just be in the way. “
However, she remains calm about AI and says that, as this level of robotic intelligence will not be around for another 50 years or so, we have sufficient time to fine-tune them and make sure that they cater to us, which will change the world for the better.
Ms Krakovna said: “If and when we build human level super intelligent AI, it would transform the world very radically.
“It would be a transformation on a scale of the industrial revolution, or maybe even more so.
“People should not be scared, but people should be paying attention.”