Technology is developing fast and artificial intelligence is no longer only a topic for science fiction. While strong artificial intelligence (simulating the human brain) still has a long way to go, weak A.I. (problem solving algorithms) is already in use.
By finding and combining information, computers can form conclusions and give solutions to common problems. Examples include different types of recognition software, chess simulators, medical diagnosis software and robots. These algorithms are able to learn and become more efficient with additional data.
As these examples show, A.I. can be very useful as it has access to more data than the human brain and can access it more quickly. So, a computer can check all possible languages and translate an old manuscript in an unknown language within mere hours while humans would take forever to accomplish the same task. Voice or handwriting recognition is very practical for transferring information into digital formats, but also to create accessibility for people with disabilities. Diagnosis software can compare a picture or a symptom with an extensive database and therefore, be more precise than a human doctor in many cases.
But there are also severe downsides. As those programmes are designed by humans, they can be compromised by hackers or abused. Some examples are hacking into self-driving cars or robots causing accidents. There have also been reports about a face recognition software that was able to predict people’s sexuality. This might seem like a very minor application, but given there are countries that outlaw homosexuality, it is a frightening possibility.
The problem lies in different types of intelligence. While A.I. is in many cases more effective than humans when it comes to cognitive intelligence (gaining knowledge and making conclusions), it lacks in emotional and social intelligence. Computers don’t have feelings or a conscience, even if sci-fi works like Wall-E or Blade Runner might want to differ. This could explain why we don’t fully trust computers.
So, algorithms are neither good or bad. They can be helpful, but also abused. They can’t be evil since they don’t have a concept of morality, but that means their “moral compass” depends on the person in control of the algorithm. And, as we should all be aware, humans can be incredibly cruel – or stupid (like orange men with red buttons), which might be just as dangerous and terrifying.