“Since you’re now working for Jeff Bezos, can you tell me how long it’ll take for the Amazon robots to take over the world?” My friend jokingly said to me upon my announcement that I will be working as a product lead on the AWS artificial intelligence (AI) team over the summer.
This was not the first time I’ve personally heard these types of jokes. Movies such as “2001: A Space Odyssey” and “Avengers: Age of Ultron” have portrayed AI systems that have exceeded the control of humans. Computer scientist Stuart Russell defines AI superintelligence as “an intellect that is much smarter than the best human brains in particularly every field, including scientific creativity, general wisdom and social skills”. AI has had its flashy moments in the last few years, from playing master-level chess to playing jeopardy.
However, I believe that the fear of AI is deeply rooted in the human belief that we are superior to all other living species and that we do not like things that we may not be able to control. As engineers push the boundaries of what AI can do, an increasing number of people have voiced their concerns about the threat a super intelligent artificial intelligence system can post to humanity. The visionary Elon Musk has expressed concern about people’s exposure to cutting-edge AI by labeling it as “our biggest existential threat”. Even Stephen Hawking said that “the development of full artificial intelligence could spell the end of the human race”.
It is clear that many people are fearful of the rapid progress in machine intelligence, but the truth is advanced artificial intelligence applications will drastically enhance human life and can easily be overcome by humans. People seem to forget that humans themselves are growing smarter alongside machines. In today’s world, a machine is made to be as intelligent as it needs to be in order to achieve the human objective. However, new advancements have people scared that a machine can have pure intelligence unbounded by humans.
I may be biased but as an engineer myself, I do not believe humans are close to developing the perfect super intelligent and self-learning algorithm. That is because human intelligence is almost impossible to reproduce. It is a complex hodgepodge of emotions and desires that are not easily replicated by a machine. Data also remains the cornerstone of AI is built. In order for a machine to achieve Artificial General Intelligence (AGI), the amount of data that needs to be used to train such a machine is beyond my comprehension. The algorithms behind AI also operates within boundaries posed by the creators and there is no evidence that there has been significant progress made to help AI reach full consciousness.
If we do get to a point where the super intelligent algorithm plots to take over the world, there will be mechanisms set in place to thwart that threat. I believe most of the fear stems from a few areas. First, that AI will steal jobs. Second, that AI will harm people.
I do not believe AI will steal jobs. Yes, there are statistics that state that 80% of jobs today can theoretically be done by machines. However, in most situations, AI is still nowhere near trusted to operate in a fully autonomous mode. My view is that AI will help guide humanity towards an age of advanced technologies where new categories of job are created along the way.
While I cannot fully dispel that criminals will not use AI to cause harm, I also fully expect governments to use AI systems to smartly enhance a country’s defenses. I believe that the fear that AI will help criminals take over the world largely stems from movies that shed a negative image on super machines. It’s mind boggling to think that the number of mobile devices on earth has already surpassed that of humans since 2014. Shouldn’t the AI apocalypse have happened by now then?
It is clear to me that many people fear AI because they themselves do not have a good understanding of what it is. Many people including my own friends do not want machines to get too smart because they fear that machines cannot be controlled. I, however, believe that most of today’s fears about AI are unfounded. Ask yourself this question. How often does Siri, Google Assistant or Alexa get your simple request wrong? The world’s largest companies cannot even produce a voice assistant that does not get easily thrown off by something slightly different. The negative rhetoric surrounding the cutting edge research will only slow innovation down. I for one am optimistic and hopeful for the future of AI. Onwards and upwards.