I’ve been spending a lot of time working out how to support the increased use of software robots in my current role. What is consistently not surprising is the view that robots come packed with intelligent and a view to taking over the world, just like the [insert traditional bad guys of your culture here]. Whichever way I try to take these conversations, the first response is invariably something along the lines of “…but what if the robot decides to do something naughty?”.
While this point of view may bring a few wry smiles and the odd eye-roll, it’s not very helpful at a time when there needs to be some, at least vaguely serious, discourse on the nature of robots as a legitimate part of the workforce. Why are software robots different to hardware robots?
Hardware Costs More
A hardware robot, generally, is expensive to purchase, costly to run and maintain, and has an overall high level of investment. There is a reason why production lines require serious levels of investment, from the simplest to the most complex. The owner can ‘t escape the financial realities of using hardware robots as replacements for humans doing the same set of tasks. There is a clear, albeit, complicated calculation for the return on investment, that allow a business owner to stack the two up against each other and come to a decision on whether one is financially better than the other – clearly, some industrialists and business owners are more ethically and morally motivated, and this is a whole super-sized can of worms that we’ll look at another time.
Software robots are cheap, at least in simple terms. The software to run them can be free for non-commercial use, and is relatively cheap even when scaled for the enterprise. The computing requirements are also low-cost, and with the proliferation of cloud computing, simple to operate. Development of automated processes, while not quite as straight-forward as advertised in many cases, is very accessible, even for non-programmers, which in turn democratises the technology. The return on investment story is compelling, so compelling that “no-brainer” is a term often associated with software robot development.
There is another difference, one that gets to the heart of our primitive fears. It’s one of competition, and when the two options are not operating on a level cost playing field, this is only exacerbated.
Primitive Fears?
Hardware automation is something we’re more comfortable with, possibly because we’re just used to seeing it in action, and can observe there are clear differences between, say, a robot arm on a car production line swinging in the next just-in-time driver-side door unit and slotting it onto the hinges with absolute precision, versus a crew of workers hauling the same door unit onto the chassis and holding it steady while another crew fix the two together. My point is that there are sufficient inherent characteristic differences that allow us to accept the need for hardware robots.
Software robots, by contrast, are stealthy, working in the shadows, performing tasks in a controlled and repeatable manner, but performing tasks that can easily be completed by a human. Perhaps not as consistently, nor for as long, but the software robot is essentially replicating human actions. The differentiators are not sufficient for us to see the need in quite the same way.
The software robot is in direct competition, and has a number of clear benefits to the employer – cheaper, simpler, more hard working, doesn’t ask for maternity, never calls in sick, has no need for a career break, and doesn’t ask for a promotion. To a short-termist employer this is that “no-brainer” solution I mentioned. Unfortunately the short-term planner will be in for a longer term headache.