From purchasing to hiring robots
Why HR should be involved in purchasing robots for your company
Take a second and think about someone you’ve worked with that was toxic to your team. Maybe they were Machiavellian, narcissistic, volatile, or just plain lazy. How did that effect your work, or the work of your team?
It is just as likely you’ve also worked with someone that elevated the team and made everyone better. In either situation it’s clear how a single person’s behavior has ripple effects across an entire group — for better or worse.
This is one reason why we have an interview process before hiring someone. It’s rare for anyone to rely exclusively on a resume or skill test for a hiring decision. When we interview a candidate for a job opening, we want to understand more than if they can do the job. We want to understand fit — both from their perspective and ours. How might they interact with the team? Will they help raise the performance of the team or are they just focused on their own advancement?
Do purchasers of robots seek the same understanding before ‘hiring’ one for a job?
At first glance you may think, who cares? Robots are industrial machinery. We buy them primarily because they are NOT human. They perform exactly as programmed and do it tirelessly, without health costs, vacations, or union backing.
It’s easy to think that robots are just replacing human labor, but in many cases the robots are also meant to help their human co-workers become more efficient. Doing so requires robots to interact with humans.
Robots cease to be strictly a piece of industrial machinery once they interact with humans
Back in the ’90s Clifford Nass, a professor at Stanford, observed that people attribute intelligence and social qualities to technology.
“People subconsciously treat technology socially. They attribute to it a rational mind similar to their own, capable of making decisions and interacting in traditional human ways.”
If you’ve ever called your GPS ‘him’ or ‘her’, or named your Roomba you’ve experienced the same phenomena. In fact, it’s the same phenomena that often causes us to view robots as unintelligent. Robots and AGVs are often perceived as slow and dumb because they rarely present enough social information for us to interpret their intentions.
A robot who doesn’t ‘fit in’ undermines human performance and the robot’s own efficiency
When robots are introduced into a work environment with humans, they become part of the social fabric of the group — much like inserting a new coworker. I won’t dig into the plethora of research work on the social and emotional aspects of human-robot interactions (but I do recommend checking out Heather Knight’s work if you want to learn more). It suffices to say there are countless factors to consider when evaluating the effect of introducing robots into a work environment. For now let’s zoom in on the second highest motivation (after reducing labor costs) for purchasing a robot: efficiency.
Robots are programed to perform tasks in the most efficient and productive manner possible. Time is money. Optimizing robots for efficiency has a direct relationship to the cost of operating. This all works just fine until you introduce humans into the equation. When a robot doesn’t perform in a socially predictable way it can undermine human performance, but also it’s own ability to be efficient in its task.
David Lu and William Smart have written extensively about understanding human-robot interactions. In one paper, they explored a very simple, yet common situation where robots and human interact: passing each other in hallways or aisles.
“When two people pass each other, they have a shared body of implicit knowledge about social situations, and swap a multitude of subtle social cues in order to manage the interaction. Both of these are typically missing in the human-robot setting.”
It’s not really the robot’s fault though. Not only are humans obstacles to be avoided, but we also move and adjust our actions in relation to how the robot moves. This causes awkward moments and can slow what could otherwise be an efficient passing for both humans and the robot.
You might think that these kinds of moments can be fixed with a fancy algorithm, but it’s not really an issue of technical challenge. It’s about managing perceptions. In Lu and Smart’s work, the solution was simply for the robot to ‘look’ at the person briefly, then move to the right side of the hall. This gave the humans confidence that they had been ‘seen.’ Though the robot’s movement to the right side of the hall was technically suboptimal, by conforming to social norms it and the human were able to optimize their collective efficiency.
Shifting from purchasing to hiring robots
The example in Lu and Smart’s work is just one of many situations where humans and robots need to interact. Two important questions for purchasers of robots are, how will the robots interact and how will we determine fit?
Hire for social awareness
You wouldn’t hire someone who places their own performance above the good of the company. Why would you hire a robot that does the same? Robots that exhibit awareness of the social cues humans use to get work done are going to interact better with their human coworkers. One of the most important social cues is showing intent.
Researcher Leila Takayama found when a robot used animations to signal intent and its ‘thought’ process, the human participant felt more goodwill toward the robot, even though it was still blocking the hallway. It’s similar to how we appreciate a driver using their turn signal. It shows they acknowledge our desire to know their intent so that we can act accordingly.
To hire for social awareness, look for robots who signal their intentions and make their internal work visible in a way human coworkers can interpret naturally and efficiently.
Hire for organizational fit
Bilge Mutlu and Jodi Forlizzi of Carnegie Mellon University studied the effects of robots in hospital environments and found that even within the same hospital, the acceptance of a robot varied widely depending on the political, environmental, social and workflow of a particular unit. It’s arguably easier for a company to change their workflow when introducing robots than it is to change their political, environmental and cultural factors. This is how organizational fit becomes a factor.
Once you articulate your organization’s unique culture and environment, you can consciously hire for fit. If there are many tight social relationships, robots that interrupt these relationships will be tolerated less — decreasing their effectiveness. If the pace of work is hectic, even the most efficient robot may prevent their human coworkers from being efficient. If the environment is clear and open, you may look for a faster robot and need fewer social features.
Hire for emotional connection
This might feel counterintuitive, but several researchers have found that performance is amplified in situations where humans feel positive emotion toward the technology (Kwan, Mutlu, and Sinno are just a few). In the most simple of examples, even a hand-drawn smile can make an impact on how well humans collaborate with a robot to achieve its goal.
This is especially important when the robot inevitably makes mistakes. When a robot inspires affection, its mistakes are more accepted by human peers. In fact, the developers of Jibo made “owning mistakes” part of the design. They recognized that:
Our imperfections make us unique and interesting to others. And, they make it much easier for others to relate us. Being vulnerable, which naturally comes with imperfections, is an opportunity for empathy and emotional connection.
When you’re hiring a robot, does it inspire an emotional connection? Does it own up to it’s mistakes?
The presence of robots, automated guided vehicles, and artificial intelligence is only going to increase in the future. We have a choice to make when introducing these technologies into work environments like hospitals, warehouses and manufacturing: Aquire them like any other piece of industrial machinery, or recognize that they are part of a social fabric and hire them to fit our unique organization. How we make that choice will inevitably affect how well robots, humans, and the entire organization performs.