Can artificial superintelligence exist, and if so, will humans become extinct as a result?
The massive growth in technology has undoubtedly impacted the lives of billions over the past few decades. We can easily contact people across the world, cars can drive themselves, and the upcoming iPhone X allows users to animate their emojis. Research surrounding artificial intelligence has similarly experienced an immense amount of progress; computers can already “understand” human speech, reply accordingly, and guess your next response. However, what if in the future, these computers become as smart, or even smarter than us? What if these robots discover a way to program and improve themselves until they surpass our limited human intelligence? What if, in a world where computers have superior intellect and capacity, humans become obsolete? Could the “singularity” exist?
In order to form a reasonable prediction about the possibility of this sort of future, we can try relying on various measurements of technological growth. One popular theory – Moore’s Law, determined in 1965 by Intel founder Gordon Moore – states that as the size of computer chips decreases, computer power doubles every two years at the same cost. Although this law appeared legitimate for decades, in recent years the theory has fallen apart due to the physical limits reached and expenses created in decreasing chip size. However, despite the demise of Moore’s Law, researchers have discovered alternative methods and materials to the traditional computer chip, even exploring the possibility of quantum computing. In other words, even though Moore’s Law may no longer be valid in predicting future computational power, the theory’s general conclusion about continuing technological growth may not be far off.
Furthermore, the failure of Moore’s Law demonstrates that as our technology changes, so will perhaps our methods of predicting that change. In fact, given that many modern prediction systems depend on technology – machine learning, computer simulations, etc – how can trust current predictions about technological advancement when that advancement will improve how we predict in the future? While theories like Moore’s Law may prove helpful in the short-term, is there any way we can rely on long-term predictions since our methods of prediction will constantly change as technology does?
There may be no way to accurately predict the existence of the singularity; regardless, wondering about the possibility is pretty fun and may just be the best we can do (for now).
“After Moore's Law | Technology Quarterly.” The Economist, The Economist Newspaper, 25 Feb. 2016, www.economist.com/technology-quarterly/2016-03-12/after-moores-law.