Yesterday I talked a little bit about the future of employment. Apparently this isn't the most interesting topic. However, it's important. The Slate series ends with some startling research that shows even scientists could eventually be replaced. I think we are a long way from those things happening. In my opinion the first things that machines will do in R&D is replace humans in the creation of incremental innovations. In fact, to some extent computers already do replace humans in some of these things. Computers do a great deal of CPU, DRAM and Flash designing. Typically, these are incremental innovations. They are building on a current technology and making improvements. Humans are required for the radical innovations, such as a new chip set, calculation methodology or what have you.
Even some advanced R&D work could easily be improved by computers. Researchers have to read a great deal of papers to keep up with the state of the art in research. As the slate series points out, this is a form of data mining and lawyers are currently using automated programs to find specific words. There's actually a branch of Science and Technology Studies that focuses on word analysis. They use similar programs and dump a few papers into it and figure out what verbal connections between the papers exist. This is a way of creating maps of knowledge. You are able to see through citations and similar word usage that a specific theory is prevalent or not. How would this apply to R&D? You could put in the materials that you're using the problems you're seeing and a bunch of papers that might be related and see what comes out. It could give you new materials new designs things of this nature. For this to work though, it's a ways away.
What does mean in the long run? That no position is safe. I don't think this will happen in our life time though. People are much too conservative to leave everything to computers. They just simply won't be accepted. Even by our generation there's too much distrust. It's going to take one or two more generations for there to be enough trust in computing and technology to allow more control to shift to them. Sure some companies will be on the cutting edge with accepting these changes, others will be laggers.
If computers can do everything why do we need any jobs, isn't the guy from CNN is right? I disagree. People will always want to work. People need to work. I'm not saying this because I'm hoping there won't be a robotic take over or anything, but because people will not allow it to happen. In general people like to feel in control. Even if you aren't the bus driver, knowing that it's a person that you can relate to makes you feel like your more in control. Leaving everything to computers requires a level of surrender. Many people will simply refuse to give up that level of control. We won't have fancy automatically driving cars for this very reason. People love to feel in control of where they go. It doesn't matter if they would be safer, save money and get places faster. They would rail against the change because they loose control.
Would we leave the future of our economy in the hands of machines? You could argue that some companies already have. For instance take the May flash crash on Wall street. This has been attributed to high frequency trading following logical algorithms, it wiped about $1 trillion in wealth, most of it was restored.
In much of my research on academic spin-offs and technology incubators there is an important component related to tacit knowledge. Know how of the inventor of a technology. This is something that we'd lose if all of our work was robotized. There's no difference in that than outsourcing. In developmental economics and innovation theories the ability to create copycat technologies is a precursor to developing their own technologies in that field. I think this is something we must keep in mind when discussing the reality of full automation. Without tacit knowledge and hands on experience with the devices and machines building the product it's very difficult to develop improvements on either.
I think that we'll have many legacy jobs hanging around for a long time. Simply because we need them to continue growing economically. Otherwise, we'll stagnate and keep producing the same technologies.