It is dead cheap compared to electronics, but they die, so you have to replace them periodically. There is the potential for reconnaissance that we are not able to do right now. That's what I'm being funded to work on for the U.S. Air Force. You don't have to send a Marine over the hill, but unarmed drones. If at least one gets back, we've got the reconnaissance data and we didn't lose a human.
Or it could fly along in advance of the convoy looking at the road and detecting improvised explosive devices. When it detects one, maybe they just fly into it and blow it up. I'm fine with that as long as they're cheap enough. We need to have that rather than a Humvee runs over the IED and we have Marines killed. That's something you should be seeing in the next few years.
How could that insect–robot hybrid work be scaled up to humans?
The hybrid there is to connect the brain, spinal cord and existing nerve endings of an amputated limb and use all that information to guide a robot system. For example, there are electrodes implanted in the pectoral muscles of the chest that connect to a robotic limb. You can use the pectoral muscles to control the robotic arm with pretty good facility.
In 50 to 75 years if you're going to be a high-end fighter pilot, you will have a neural implant that connects directly to the aircraft. Rather than using your limbs to control the fighter you will control it directly with your brain. That will give you reflexes that are superhuman. Once you can do that, it's not that big a deal to connect to an exoskeleton.
Is it more likely we'll just have robots fight for us?
That depends: When will we be able to build a machine that is as intelligent as a human? When can we send a robot to Elysium and have it do the task that Matt Damon does in this movie? A truly intelligent robot, one as intelligent as a human being—we are a very long way off of that. We are really clueless about cognition.
Even in insect brains we have no clue about the basis of decision-making. In humans it is so complex. So maybe focus on how insects make decisions. Maybe it's a scaled down version of how a rat does it, so scale it up to a rat, then a cat. But then humans, with language, have extra tricks.
For the foreseeable future, even in 2254, you're going to want a human being in the decision loop. So you will augment a human being rather than create a truly autonomous device.
I can tell you for a fact that the U.S. military leadership is very interested in keeping humans in the loop. The military has access to autonomous vehicles that they don't want to use. They are actually very conservative and concerned about things like civilian casualties. They want a human to decide every time a trigger is pulled. Humans make mistakes, too, but they can be forgiven. When an artificial device makes a deadly mistake, it cannot be forgiven. They have to be utterly reliable.