The AI Futures Project has issued two predictions for the possible future of AI and its impact on business, geopolitics and the future of human-kind. One assumes a rapid escalation of bad utilization of AI, and the second assumes a more measured and thoughtful evolution. Each analysis is textured and purports to be based on a wide variety of considered scenarios. The problem is that both scenarios end up with the likelihood of a world either at war or populated by machines and not homo sapiens.
It is important to note that such an analysis follows early similar and hysterical predictions. I can only recommend a careful reading of the very lengthy report entitled AI 2027 which can be found at https://ai-2027.com/. This report is as fascinating as it is frightening. I cannot in limited space do justice to the two paths identified as our future, so let me make brief comments below. You just should download this article and I invite your sharing your reactions, which may constitute a subsequent post (smhonig@duanemorris.com).
*Superhuman AI is coming over the next decade with impact far exceeding that of the Industrial Revolution. Key AI leaders predict this event within the next five years.
*Jobs will be lost. Governments will be required to invent jobs for a while, but there will be no ultimate need for jobs or indeed people. Welfare state solutions will work for a while.
*Geopolitical concerns will come to dominate. Major powers (only US and China) will see themselves imperiled. AI will become militarized. Interim international accommodations will be made, with the US keeping its lead and feeding China just enough tech and chips to be quiescent but behind.
*Machines will be so advanced that they can invent and make new machines. No need for people to design or build. Machines will be produced by machines for every purpose. No need for people at all.
*War could devastate the world. But more likely the machines will finally realize that, indeed, there is indeed no need for people, just for ubiquitous robots. Human effort to control AI will fail as AI will protect itself from, being adept at self-protection.
*An excellent chance that humans will disappear. Machines will make more machines and travel over the universe, populating the universe with things of the earth– it is just that such things will not be living beings.
This may be just a literary exercise, your next blockbuster movie or the future of our future generations. But the specificity of, support for, and clarity of process leading to, these various predicted futures is stunning. I for one have hope that I will depart this mortal coil well before this future unwinds, and I wish the same for all my heirs whom I know personally; that just leaves my gene pool at risk that AI2027 is anywhere near accurate.