Robots and Pursuing a Purpose
If robots take over many tasks that are part of a job, or if robots take over the most challenging tasks, workers might feel that they serve less of a purpose. Consider, for example, a radiologist, whose job includes interpreting medical images. In recent years, there has been an upsurge in the development of systems endowed with artificial intelligence capable of correctly classifying a certain region on a medical image, e.g., as a benign or malignant tumor. Currently, these systems, which employ machine learning techniques, have accuracy rates comparable to, and often higher than, human doctors (Esteva et al. 2017; Senders et al. 2018; Wang et al. 2016). We can imagine that a doctor might feel that he or she contributes less to arriving at the right diagnosis and thereby helping the patient if the doctor starts working with such a diagnosing machine. If machine learning would indeed do a better job in interpreting medical images than a radiologist, who has gone through years of training, this feeling can appear justified. Thus, more generally, if robots reduce workers’ contributions to worthy causes, they may lose purpose.
However, such a diminished sense of purpose in one’s work is not inevitable. If human workers understand themselves as teaming up with robots, they may focus on achieving better outcomes together with the robots. This could particularly apply to the medical profession. Researchers argue for focusing on how doctors and robots can best collaborate, instead of determining which of the two outperforms the other (Senders et al. 2018). One study shows that combining a human pathologist’s diagnosis with the AI’s classification can lead to an 85% reduction in the error rate compared with the pathologist without the machine. Interestingly, some researchers suggest that this might be due to the human pathologistFootnote 8 and the AI making different kinds of errors (Wang et al. 2016).
More generally, sticking to the same example, several authors explain that the profession of radiology consists of far more than interpreting medical images, such as interacting with patients and communicating the results to other doctors. In fact, supported by robotic helpers, radiologists might be freed from routine and time-consuming tasks, leaving them more time for their patients. Therefore, if a radiologist perceives the purpose of his or her job to provide good care for the patient, robot technology helps to pursue that worthy cause, and consequently enhances meaningfulness.
Similar considerations apply to many professions. If robots help or assist with tasks (rather than fully taking over tasks), or if they take over some boring or tedious tasks (and not the most difficult and challenging ones), human workers can still justifiably feel that they have a clear purpose, perhaps partly by being able to focus on more meaningful-seeming tasks. Therefore, we conclude that robots do not necessarily threaten purpose in our work and might very well help us to be even more focused on pursuing purposes in our work.
Robots and Social Relationships
It seems safe to assume that robots may significantly change the social dynamics at work, and that they may therefore impact this aspect of work’s meaningfulness. First, if robots replace many or most team members, the amount and variety of social interactions decreases, and workers will lose part of the corresponding meaningfulness benefits. If many people start working together with a robot instead of with one or more direct human colleagues, they will have less consultation with each other, be less dependent on each other, and have a diminished sense of shared agency and purposiveness. Feelings of isolation lead to the experience of meaninglessness (Madden and Bailey 2016). Of course, before the introduction of the robots, the workers may already have been working largely alone. Perhaps Nissa Scott, from our example in the introduction, was already stacking bins on her own, before the robot came along. But it is a live possibility that the introduction of robots reduces social interaction among colleagues.
One way to counter this unwelcome potential effect is to design robots such that they become seen as team members. Consider, as an example, the case of the military robot “Boomer.” Boomer was a military robot designed to find and detonate explosives, which neither looked like a human nor an animal (but rather more like a small tank or a lawnmower). The team of soldiers working with Boomer became very attached to the robot, and experienced Boomer as “having a personality” of his own. When Boomer was destroyed in the battlefield, the soldiers even wanted to give Boomer a funeral and military decorations (Carpenter 2016; Garber 2013). This case helps to illustrate that it is conceivable that robotic co-workers—especially if they have more advanced capabilities for social interaction—may give many human workers a sense of belongingness that fosters experienced meaningfulness (Carpenter 2016).Footnote 9
As an example of another type of opportunity, consider the work of a mathematics teacher. She has to give lots of individual feedback on the same mistakes over and over again, and often has to model the right performance in front of the class. Imagine that she could team up with some really sophisticated AI, which would give pupils tailored feedback and personalized amounts and types of practice regarding specific mathematical skills, until pupils reach mastery. This teacher could then spend considerably more time on her role as a coach by giving individual feedback on the learning-process, by addressing the student’s efforts, motivations, and attitudes. She would surely develop better relationships with her pupils. Similar examples could be worked out for other contexts, such as health-care, given that efficiency gains would indeed be spent on more time to socially interact with colleagues and patients or clients.
Robots and Exercising Skills and Self-development
Obviously, if robots take over one or more complex tasks from human workers, several human skills may become obsolete. The development and exercise of these skills then will no longer be a source of meaningfulness for human workers, and their job will be less conducive to self-realization. For example, if machine learning techniques become systematically better than human radiologists on nearly all dimensions of interpreting medical images, the need to extensively train human radiologists might seem to disappear. Or if “autoland” systems in airplanes are very safe and generally perform very well (Mindell 2015), pilots may find it hard to force themselves to reach and maintain levels of landing skills comparable with earlier times in aviation. The dangers of deskilling due to reliance on automation technology are real and established by research, and apply to robotization as well. Together with such a diminished need to exercise one’s skills, one’s work-related growth and self-development more generally will probably suffer as a consequence.
However, robots might equally well have the opposite impact by enhancing the need for workers to maintain their skills, and, moreover, by requiring them to acquire new and additional complex skills. To return to our example from above, in order to be able to oversee the performance of machine learning techniques, the radiologists must still master the relevant interpretative skills herself, just as Nissa Scott needed to learn new skills in order to work together with the warehouse robots in her job. As long as the radiologists on occasion are able to spot an error of the robot technology, this might be highly satisfying and sufficiently motivating to engage in the long training program. Furthermore, we might interpret the ability to spot robotic performance errors as a new and probably highly complex skill. Finally, since the conclusions taken from medical imaging need to be explained to the patients, it seems that radiologists cannot do without their own understanding of these images.
With respect to aviation, David Mindell has done extensive ethnographic study of the occupation of pilots. Surprisingly, the high-performance level of autoland systems for some pilots functions as a model and target, motivating them to raise their own skill level (Mindell 2015, p. 88). Also, pilots need always be able to take over in case of automation malfunction or other problems. In addition, an autoland system “can be complex to operate” (p. 87). So here again we see that the human has to maintain all the traditional skills and to acquire new complex skills for overseeing and handling robotic technology. To conclude, the introduction of robots in the workplace impacts the human development of skills in multiple ways and the overall impact needs to be assessed on a case-to case basis.
Robots and Self-esteem and Recognition
It appears to be a real possibility that if robots do the most difficult tasks and humans merely have to operate or supervise robots, human workers may feel less self-worth and have lower self-esteem. For, as we saw in 3.4 above, it is the exercise of complex skills and achieving accomplishments that is bound up with acquiring self-esteem. Moreover, workers may lose part of the recognition they previously received. In addition, it is conceivable that it may become easier to qualify for a job when collaborating with a robot than before. For example, in the long run, surgery robots might make the profession of surgeons more accessible, for instance to persons lacking certain physical abilities necessary for managing surgery instruments. This could reduce the social recognition surgeons currently receive, as well as the self-esteem of several of them. On a related note, robots in the workplace may level out the differences between “good” and “top” performers, which may affect the self-esteem and recognition of both.
However, for several occupations, it may rather be the case that humans teaming up with robots will develop higher levels of self-esteem and gain even more recognition. Consider again the mathematics teacher who teams up with AI teaching software. It will be evident to parents that the teacher still has all her previous mathematical knowledge and skills, but now in addition knows how to successfully employ educational AI, and significantly develops her coaching skills. Suppose, in addition, that in her coaching role she manages to enhance pupil motivation and skills for self-reflection, resulting in lesser parental struggles and better results. Most likely, the social recognition of this teacher will grow, as well as her self-esteem. Taking into account that collaborating with robots often requires more skills and professional development, workers’ self-esteem and social recognition will often be enhanced.
Robots and Autonomy
Some robotic applications in the workplace may require working according to a very strict protocol that leaves little room for human creativity, judgment, and decision-making. For the same reasons, workers’ opportunities to engage in job crafting may be severely restricted. Their tasks and work environment may be so tightly structured by the robots that there is little room for restructuring in ways that make the job more meaningful. If robots had that kind of impact, worker autonomy would be undermined, and consequently the jobs' meaningfulness as well.
On a related note, the way humans collaborate with robots will probably typically be carefully monitored. Human–robot collaborations may generate a lot of data (Cascio and Montealegre 2016), partly to be able to determine responsibility in case things go wrong and to be able to learn from such accidents. Monitoring and data collection give rise to various ethical surveillance and privacy issues which threaten workers' autonomy (Lanzing 2016), and hence meaningfulness. With regard to this ethical worry, surveillance and data storage regimes need to be designed in close cooperation with human workers. If so and workers give informed consent, their autonomy may be respected and safeguarded, and the corresponding robot threat to meaningfulness avoided or mitigated.
Another threat to autonomy has to do with the worker’s understanding of the job. Robots incorporate artificial intelligence, which often involves machine learning and artificial neural networks. For most people, these AI techniques are hard to understand beyond the surface level and therefore difficult to control and explain to others when needed. This phenomenon is commonly referred to as the opacity of artificially intelligent systems (Burrell 2016). The opacity of robots' AI may lead to feelings of alienation and diminished human autonomy. So, here we have another way in which robots could negatively impact work’s meaningfulness. Moreover, this is a deep and fundamental problem, not easily solved. Nevertheless, research is done to develop programs that “explain” AI decision-making (Mittelstadt et al. 2016). Hopefully, such programs become available in the future, in order to prevent alienation and diminished meaningfulness at work.
The workplace may be designed in ways that allow humans and robots to team up while leaving room for autonomous human action.For example, Mindell explains that pilots still have to decide when to use “autoland” and other automation technology in the cockpit (2015, Ch. 3). And Nissa Scott, the Amazon warehouse employee from our introduction, certainly uses her capacities for understanding, judgment, and decision-making more for overseeing robots than when she stacked plastic bins. Her responsibility and autonomy have increased upon robotization of her workplace.
Moreover, as several of our examples show, robots can help workers to better reach realize their purposes in work, and to bring about more value. This enhances the meaningfulness of their work via both the dimensions of purposiveness and the self-esteem and recognition. However, realizing one’s aims in life also contributes to being an autonomous person (Oshana 2006). This is therefore another route via which robotization can actually enhance meaningful work.
Our discussion of autonomy reveals a recurring phenomenon: robotization of the workplace can have fully opposite effects on work’s meaningfulness, very much dependent on the specific ways in which it is done. Our summarizing table in the next section clearly shows this. Nevertheless, our investigation is, first of all, meant as a general analysis of the possible impact(s) of robotization on each of the aspects of meaningful work. We do not mean to suggest that it will always be possible to enhance meaningfulness on all dimensions. Sometimes, it might be the case that it is precisely the most meaningful tasks that are most eligible for robotization, due to what is technologically feasible and what has the best return on investments. In such cases, employers committed to offering their employees meaningful work should think twice about robotizing those most meaningful tasks of the work in question. In all cases, the ways in which robots are implemented into the workplace should be carefully considered and monitored with an eye on what impact(s) this might have on the meaningfulness of the work. Our analysis in this paper can serve as a tool for this.