If you follow the commentary on artificial intelligence, then you’ll be aware of a certain tension. On one hand, some people are saying AI will change everything. On the other hand, some people are saying the hype is overblown.
Many who say the hype is overblown are AI experts. In fact, many of these experts who would rather you not use the term AI at all and speak instead about the current tool: machine learning.
The heart of the issue is that AI is a big deal, but it may not be quite the big deal it’s made out to be. Allstate CEO Tom Wilson has said of AI, “It’s going to rip through this economy like a tsunami.” This is true, but it is true in the same sense that the internet ripped through the economy and smartphones ripped through the economy and global supply chains ripped through the economy.
Mark Edgar, the founder of The Goat Rodeo Project, says AI should be taken to mean “Another Innovation.” That is, it’s a big disruptive innovation, but we’ve been living through big disruptive innovations for 100 years. We know how to handle them — or at least, survive through them.
For the next 20 or so years, AI will continue to be a disruptive innovation, much the same way technology of all kinds had been disruptive in our lives and the economy. We should treat it as such. Things may look different. And at some point, maybe in 20 years, maybe in 200 years, machines will have near-human intelligence, perhaps even super-intelligence. At this point AI will not be just another disruptive innovation, it will be a new thing.
Robert Miles, a Ph.D. student at the University of Nottingham, cites the analogy of an island facing gradually rising sea levels. In the past, whenever this happened, people just moved up a bit higher on the island, so it wasn’t a real problem. However, it will reach a point where sea levels rise over the highest point of the island and then there will be nowhere to go. Similarly, with smart technology, we’ve been adapting easily because there were always higher-level tasks that machines couldn’t do. When machines begin to approach human intelligence then there will be no place left for humans to go.
It’s a good time to start worrying about how to govern AI so that we will be ready to face human-level AI. Until then, it’s just another innovation, crucially important, but not something radically different than what we’ve faced before.