by Lucine Avanesian
Will artificial intelligence ever be able to think like a human? The rapid advancement of AI, along with our natural mistrust of robots, has created fears that it will one day develop uncontrollably, leading it to take over the planet and wipe out humans if it thinks we are an impediment to its aims. The “AI singularity” is a term used to describe this point in time.
What Is Quantum Theory & What Does It Suggest?
One argument against the existence of such an unstoppable, ever increasing intelligence is that it would have to be able to accurately anticipate the future by definition. Quantum theory, one of modern science's most important techniques of describing the cosmos, claims that predicting the future is impossible due to the randomness of the universe. But what if we believe it is impossible to anticipate the future because we aren't sophisticated enough to know otherwise?
Intelligence is a difficult and enigmatic notion with no universally accepted definition. However, there is widespread agreement on some of the elements that make up all forms of intelligence. One of them is problem-solving skills, which necessitates the ability to plan by anticipating the future. Understanding existing conditions, predicting how the environment will evolve, and anticipating the effect of the actions that will be taken are all critical steps in solving an issue.
The Random Universe
According to recent physics hypotheses, the universe is chaotic and random. Take, for example, unstable chemical elements that eventually decay into another substance through radioactive decay. You can guess how long a specific amount of this element will take to decay, but you can't predict when any one atom will. Similarly, you can measure a particle's position or momentum, but for quantum-theoretical reasons, you can't know both at the same time with complete accuracy.
If these theories are right, they suggest that the universe is ultimately unpredictable, chaotic, and unstable beyond a certain level of complexity. This would imply that any rising intelligence would eventually reach a point where it could no longer enhance its future predictions and so could not grow in intelligence any further. In other words, there is no danger of a runaway AI because the universe's physical rules impose some very strict hard restrictions. Given the recognized limits of weather forecasting, an AI system will not be able to outwit humans by planning future activities using exceptionally accurate long-term weather forecasts.
So Will It Take Over or Not?
It's reassuring to suppose that the nature of the cosmos is preventing an AI arms race in some way. However, there is another viewpoint. What if our cognitive and reasoning powers are so restricted that we experience the cosmos as random and chaotic?
Taking this viewpoint, it's possible that the universe is instead deterministic, and so entirely predictable, but in a way that we as humans cannot comprehend. Albert Einstein contended that quantum theory was a partial representation of the cosmos, and that there must be hidden variables that we don't yet comprehend but which contain the key to predicting future events. The notion of an AI singularity would be thrown out the window. A super-advanced intellect might be able to uncover these hidden variables and so comprehend the universe's predictable nature, allowing the machine to reach its full potential.
On a practical level, the singularity does not appear to be a realistic possibility, considering how restricted AI is right now. Recent AI achievements have come from narrow AI, which is meant to accomplish a specific task, such as playing chess or driving a car. While limited AI can outperform people in some tasks, there's little reason to believe that more broad AI that can mimic humans' capacity to respond to a variety of tasks will be provided in the near future, putting humans at risk. However, we can't fully rule that out. Given our limited understanding of the nature of the world and the potential of AI, it may be prudent to play it safe. AI will have a significant impact on human society even if there is no singularity. We must work as hard as possible to ensure that AI benefits mankind rather than poses a threat to it.