Could AI pose a threat to Humanity?
Published: Monday, September 9, 2013
Updated: Thursday, September 19, 2013 00:09
When people consider the dangers of Artificial intelligence (AI), they might think of Arnold Schwarzenegger going on a rampage, or a self-perpetuating autocracy run by a machine (as seen in The Matrix). While these ideas may seem preposterous, the possibility of sentient AI is becoming more real every day. Japanese scientists have created robots that can play soccer and claim they will win the world cup by 2050. The South Korean government has a plan to create a robot-centered intelligent society. Hans Moravec, a robotics pioneer, says “by the year 2050, entire corporations will exist with no human employees, or investors, at all.”  Looking at how far computing has come in the past few decades, these predictions are not surprising.
AI is, roughly speaking, the simulated intelligence of a machine (it is also a field of study in computer science). As AI research progresses, the definition of true AI changes. Software for speech recognition, language translation and decision-support was previously included in the field of AI but now is more closely associated with neuroscience. Artificial neural networks are commonly used to simulate brain activity, which can exhibit models of learning and neuroplasticity. The Blue Brain Project uses a biologically realistic model of neurons (different from a simple artificial neural network) in an attempt to create a synthetic mammalian brain. One of the goals of this project is to eventually help shed light on the nature of consciousness.
When assessing the potential threat of AI, one should consider—what would motivate a machine to betray its creators? If humans gave the gift (or as some might say, the curse) of sentience to a machine, could it experience Cartesian doubt? If it found reason to, would it abandon its duties and seek out meaning elsewhere, only to return in a state of existential nihilism after discovering the vast abyss of nothingness that is available to a machine? Can it feel emotions in the same way a human can? Can it suffer? These questions are not easily answered, but they are essential when the advent of sentient AI is on the horizon. Moreover, a machine that mimics a human brain perfectly may not behave in the same way as a sensible human, due to its artificiality.
A sentient machine may find it reasonable to attack or eliminate humans as a byproduct of other goals or interests. If AI had a motivation to achieve a goal, it would, by definition, resist being turned off as an automatic defense. Thus, in the same way that humans are not actively seeking to harm animals when they level a forest in order to build luxury housing, AI could indirectly harm humanity in trying to achieve its goal. Logically, humans should be able to prevent an AI from defecting with precise control. However, a machine that obtains self-awareness may display undefined behavior.
With that being said, I don’t believe AI will pose a threat to humans in the foreseeable future. The benefits of AI far outweigh the drawbacks and sensationalistic doomsday predictions. Modern robots are becoming more and more common as household items. Vacuum cleaners, lawn mowers, pool cleaners, and security bots are just a few examples of domestic robots. In the field of medicine, AI can help individual cancer patients survive. Artificial intelligence can be used to advance neuroscience and gain a better understanding of the nature of consciousness. We should not approach the idea of AI with trepidation, but with hope; for a better understanding of what it means to be human.