Development of AI is happening in parallel to massive and disruptive demographic shifts. That’s good, Deep Learning Group founder Dr. Adrian Kaehler tells Disruptive.Asia, because AI will be the solution we need to handle that shift.
Disruptive.Asia: Let’s start with the two key mega-trends you’ll be talking about in your presentation at Convergence 05 at about the rise of AI and shifting demographics. What exactly are you seeing there?
Dr. Adrian Kaehler: On the one hand, we are seeing unprecedented advances in what has been traditionally called “artificial intelligence” or “AI”. We are seeing things like computer programs that can beat the very best human players at subtle games like Go, and that can caption images and video with an accuracy that, only a few years ago, we would have thought were a generation or more away. In parallel, in a totally different aspect of the human experience, we are seeing substantial changes in families, when people marry, when and how many children they are having, as well as how long people live. The result is changing demographics, across the entire world, in which younger people are a relatively smaller part of the population than ever before, and older people are a much larger part of the population than ever before.
How can AI solve these demographic challenges?
These changes in demographics means significant challenges in how we live our lives. On the one hand, the number of people living to such an age that they may need a lot more assistance in order to remain productive, independent, and happy is increasing rapidly, even as the number of young people who can provide that assistance is effectively declining. The challenge is to apply the new technologies that we are developing in order address these issues. This can be done in many ways, but the key observation is that AI is not “destroying jobs”, it is freeing up much needed resources.
What kinds of challenges or barriers does AI face in being able to be deal with those demographic shifts – for example, fear of AI, bad regulation, viable business models, politics, human workers refusing to adapt to whatever solutions are presented, etc?
The biggest challenge is in ordering the market in such a manner that the resources that need to go to these specific problems are actually addressed to them before the problems become too great. From my own point of view, I see a lot of fear, ranging from very legitimate fears of dystopian economics, to science-fiction type “Terminator” fears, and these things slow us down when we need to be working hard to solve real problems with these technologies.
How far along are we in terms of AI’s ability to fulfill this goal?
The hard thing is that the goal is moving. Every year, the number of older people needing assistance to live their daily lives grows relative to the rest of the population. So far, we aren’t doing so badly. For example, technologies like Lyft are wonderful for providing mobility to people who might not be in a position to drive themselves anymore. Of course, a system like Lyft only makes modest use of AI-type technologies, but someday all Lyft cars will probably be self-driving cars.
How far do we have to go?
This is a problem for demographers, but in a nutshell, we need a productivity augmentation of about +50% across the board, just to account for the labor needed to care for the future population distribution.
What’s the timescale you’re looking at in terms of how these trends will play out and how AI will solve the challenges they bring?
The timescale for these demographic changes is about the next two generations. Of course, that doesn’t mean that we have 40 years to start, it means that the effects I am talking about will onset gradually and continuously over that period. In terms of AI solving these problems, we need to start developing and deploying solutions now. Systems like Assisted Living cannot scale to accept another billion 65-year-olds, so we need to get working now on solutions that can help vast numbers of people enjoy their independent lives in a manner that our society can afford to deploy. These are massive systems, so it is probably fair to say that, even given the advances we have made, we are probably behind the curve.
Dr. Adrian Kaehler is a keynote speaker at Convergence 005 | SWITCH, which takes place September 18 at the Marina Bay Sands in Singapore. Disruptive.Asia is an official media partner for Convergence 005. For registration and other details, click right here.