Some scientists believe that Artificial Intelligence (AI) can become the great future hope and the savior of our planet. It is difficult to say if this is realistic, because at the present time AI is still in its infant stages and there are many unsolved aspects. After all we might be able to get a little glimpse into the future of AI by looking at the present tendencies and learn about some of the common challenges that AI scientists are struggling with.
Artificial Intelligence is not just one well-defined thing or system. Today there exists various types of AI which are different from each other. One type tries to implement human behavior, consciousness, feelings, emotional intelligence, and even our morality. This type could be termed humanoid robots with AI (3). As the term indicates this kind of AI is often built into robots that are modeled over known biological organisms such as human beings or animals. One of the ambitions with this kind of AI is to incorporate the ability to make moral decisions – like us.
But at the present time there are obstacles before this is possible. According to Professor Anil Seth from the University of Sussex, AI robots are at the present time very far from being able to imitate human behavior. In the video “Robots can predict the future … and so can you“, he explains that one of the fundamental problems is that humans: ”…perceive the world…with and through our bodies, emotional responses, and bodily responses”. He continues to say that these elements are ”…absolutely core to both consciousness and intelligence” (1). He concludes that ”So far AI hasn’t really incorporated bodies in this way” (timecode: 1:54-2:09). In the same video Alan Winfield, another professor from Bristol Robotics Lab agrees. Even though it is quite easy to build robots that look like human beings, their intelligence is often not higher than washing machines. He terms it ”the brain-body mismatch problem” (timecode: 2:21-2:22).
Another type of AI is sophisticated, machine-learning-systems that can learn by themselves to solve problems that we humans will never have the capacity to solve because it requires massive amounts of collected data. This type of AI is often termed Analytical AI (3). Chess-game computers and image-recognition systems belong to this category.
But there are also obstacles to this kind of AI which must be addressed before it can be let loose in the real world. If AI is focussed one hundred percent on solving a required task, then it might unintentionally break moral codes, laws or other limits that exist in a given culture. So there are drawbacks to both kinds of AI.
So there seems to be a necessity to bridge the gap between the two kinds of AI systems. If AI systems can both work with loads of data and at the same time understand and imitate human behavior then it might be the best way. That might be what professor Alin is referring to when he says: “Maybe we shouldn’t be looking to replicate human intelligence in general AI. Maybe it’s more of a complementary relationship that we should be seeking.” (02:58 – 03:07).
The great question is now how this complementary relationship could be like in the future. Right now there is one big problem and that is not the technology. According to Professor Anil Seth it is humans themselves: “Human intelligence is actually a bit of rubbish. We make irrational decisions all the time” (2:41-2:46). And Russel agrees: “People are irrational, inconsistent, weak-willed, computationally limited, heterogenous and sometimes downright evil”(2). But if this obstacle can be overcome in the future, then according to Russel AI might have a very bright future for the benefit of humanity – even the ”…eradication of disease and poverty are not unfathomable…”, he says (2).
AI is still in its infant state and we cannot really know what it will become in the future. But isn’t that just how it is with all new technology? Who could have predicted that the telegraph system one day would be exchanged with the world wide web? So let’s not be quick to judge the seemingly useless and unintelligent AI robots we have today. Maybe they have a great future. But let us keep a watchful eye on them so they don’t develop and become like us with all our irrational and evil tendencies and end up conquering the whole world.
1: “Robots can predict the future … and so can you – video” (video) Joseph Matthews, Michael Aaglund, Michael Tait and Richard Vine, The Guardian website, 2 February 2017, viewed 21 March 2017. (https://www.theguardian.com/lifeandstyle/video/2017/feb/02/robots-can-predict-the-future-and-so-can-you-video)
2: The rise of robots: forget evil AI – the real risk is far more insidious. (text) Olivia Solon, The Guardian website, 30 August 2016, viewed 21 March 2017. (https://www.theguardian.com/technology/2016/aug/30/rise-of-robots-evil-artificial-intelligence-uc-berkeley)
3: Siri, Siri in my Hand, who’s the Fairest in the Land? Kaplan Andreas; Michael Haenlein (2018)