We’re getting a good flow of articles about these sometimes scary but often cute mechanical assistants, from USA Today and the Wall Street Journal to the Financial Times and beyond. Despite the concentration in business publications, there hasn’t been as much on the profitability of the things as there has about the age-old concerns of what jobs they will replace and when, along with what they could potentially do to us as well as for us. Since they are becoming ever more common, in forms ranging from vacuuming Roombas to the human-like remote-operated “geminoid” demonstrated at March’s Austin South by Southwest Interactive Festival, it is increasingly critical that we understand the fundamentals about their nature. Here are some things we should all agree upon.
First, artificial intelligence does not mean consciousness. It is possible that robots do have that now – one school of thought on the matter sees it as coming from computation, which means that my desk calculator is aware too. Yet AI doesn’t confer that, and no matter how capable machines seem to be, they are no inherently smarter, as a corporate instructor put it to me decades ago about uncontrolled computer disk storage devices, than overhead projectors.
Second, robots need controlled environments. That is the other side of their reliance on algorithms, and why, for example, they can crunch numbers at nearly the speed of light, burying a human genius, but cannot replace hotel maids.
Third, progress on them is one-way, and we and they never regress. Some knowledge in any field is always lost, but there is simply too much work on robotics for that to be much.
Fourth, just because worry about robots taking over the majority of jobs did not happen when it was first a broad area of concern in the 1930s, or again in the 1960s, does not mean they never could. It is conceptually possible for any task with instructions that can be put in the form of “if A, do B” to be automated away. That also explains why the thought of them taking over only or even mostly manual jobs is long obsolete.
Fifth, making projections, and being concerned about and wanting to prepare for job losses, does not make anyone a technology-hating Luddite. Whatever our views are on where we are going, it is crucial for us to forecast it.
Sixth, when robots, computers, or anything else technical is used to complete work in place of humans, the number of jobs it creates is almost always buried by the number of jobs it eliminates. If that is not the case, as may have happened with office personal computers in the early 1990s, the technology is then not saving businesses any money.
Seventh, on the other side, the idea that automated systems are bad if they make jobs obsolete is a trap to avoid. We still don’t know how best to deal with robots and computers, but they are good things, and smashing them metaphorically as protestors physically did weaving machines two centuries ago is, in the long run, destructive to our prosperity, health, and happiness.
Eighth, consumer acceptance and resistance to robots will be a huge determinant, maybe the largest, in where and when they succeed. Elderly Japanese have generally adapted well to cyborg caretakers, but would older Americans? Fast-food drive-throughs now greet us with human-sounding disembodied speech, and nobody seems to mind – would that be the same if those voices were metallic? How about C-3PO-looking bartenders, who in a few years may be technically ready to run their places?
Ninth, we must still be concerned about semiautonomous robots programmed amorally enough to do more damage than help. As Stanford law fellow Vivek Wadhwa put it, artificial beings may get us to something like Star Trek, with its interstellar voyages, or to more like the post-nuclear rudimentary civilization of the Mad Max movies. Technology has always been a fine servant but a poor master, and understanding the fundamental facts about what it is and isn’t will help us choose wisely. Even if the robots haven’t taken over or made all of us obsolete, it’s not too soon to think about it.