Machines of Loving Grace, part 2

In books by Henry WolfLeave a Comment

This is the second part of this review. For my earlier thoughts click here.

This book might well have been subtitled “A Love Letter to Doug Englebart” due to his influence on the author, John Markoff. Each chapter stresses the benefits of intelligence augmentation (IA) over artificial intelligence (AI), without clearly defining a difference between the two. In the final chapter, for example, robotic caregivers are considered a type of IA because they would work with the elderly. That they would displace huge numbers of human caregivers is not mentioned. When looking at robotic machinists, on the other hand, the AI aspect is the focus. The stress on in the workers that would be displaced. More troubling is the shallow manner in which the topic is handled. Markoff could have gone into the ways in which the current human society could change, for better or worse, as a result of these new technologies. Unfortunately, he fails to see beyond the immediate job loss that could occur.

Using the example above, perhaps we do not want the elderly to be stuck with robotic caregivers. Consider a world in which plenty of people have lost their jobs to factory robots. Society could handle this by implementing a basic income. Those who lose their jobs could be retrained in any number of fields, one of which could be caregiving. Provided a basic income, people may be willing to work as caregivers even as volunteers or at least at a cost less than what is necessary to produce a robot. Perhaps they would pursue something else, but my point is that Markoff ignores the ways in which society could change with the technologies we are developing.

I could write this blog

Including a poor haiku

In every post

A related issue, that I touched on in my initial review, is that there is no clear dichotomy between IA and AI. The book leave us with the thought that it is up to the engineers building the systems to keep the human in the loop. The fact of that matter is, though, that the human may be the bottleneck. One could always build a human into any system. I could write this blog, including a poor haiku, in every post. Doing so, however, would not improve the blog in most situations.

We have generally used IA because AI has not been powerful enough. Once AI reaches a level of ability that is greater with the human out of the loop, it makes no sense including one. Why choose a car you have to drive over one you can fall asleep in and wake up at your destination? AI researchers understand that someone else will optimize if they do not. Restrictions on AI research in one location will only cause those in other locations to thrive. If these regulations are passed in California, there are plenty of other locations to test autonomous vehicles. So it is with self-regulation. If Google does not build a fully autonomous vehicle, another company will.

In short, it is not a matter of whether we pursue IA or AI, it is a matter of how society will change to accommodate the shift to a world in which humans are not the only dominant intelligences on the planet. While the book provides an interesting, yet flawed and heavily biased, history of robotics and artificial intelligence, it fails to support its thesis satisfactorily. ★★☆☆☆

Leave a Comment