MUCH WORK IN organizations today is done in teams. In an interconnected world, it is not a surprise that people often work together in virtual groups to achieve goals they cannot achieve alone. Teamwork is instrumental to achieving innovative solutions within complex environments. Teams are better able to bring a diversity of ideas to the table, respond quicker to complex challenges, and allocate ownership for tasks to multiple people at the same time. Deloitte has even suggested that in today’s digital world, organizations increasingly structure their work setting as being based on a network of teams.170
Even though volatile team dynamics are crucial for delivering fast and innovative solutions, it is also necessary for each team member to have a list of tasks they need to complete. Such tasks, which must have clearly defined criteria, can then be used to decide what kind of person is needed on the team. Today, teams require fast and transparent knowledge analysis, and therefore a new kind of member is entering the team. This new team member is not human, but algorithmic in nature.
Iron Man as the example to follow?
Inspired by the character J.A.R.V.I.S. (just a very intelligent system) in the movie Iron Man, we strive to build teams that rely on co-creation between humans and algorithms. We need to combine our strengths to cope with future challenges. A variety of tasks will thus be executed by either a human or an algorithm, but the most important goal will be to ensure that the efforts of both types of members can be integrated to create the required added value. For this effort to succeed, we need leaders to co-ordinate the efforts of humans and algorithms in ways that create something new. If humans and algorithms both represent a figure of one, then leaders need to ensure that their added value is a three. To achieve this, a leader’s task will be two-fold.
First, as discussed in chapter seven, leaders need to empower humans and algorithms in their own specific ways. Algorithms will conduct tasks that require the skill of analyzing enormous volumes of data to optimize strategy. Machine intelligence makes existing data more transparent so that it can be used to make faster and more effective decisions. Humans will be used in sense-making tasks where perception, ethical judgment, intuition, empathy, and creative thinking is necessary. These so-called soft skills are especially needed in team efforts as they help to tacitly co-ordinate work relationships.
Second, leaders will have to ensure that the dynamics of a team composed of humans and algorithms are properly managed. In other words, leaders will now have to manage the new diversity. I refer to new diversity as today’s reality, where algorithms will become the new teammates of human employees. As a consequence, teams will become more diverse, in terms of the skills on offer, but also in how the work will be done. Such a reality brings forward new challenges for leaders, the most important of which is being to understand what it is that can make interactions between humans and machines tick.
Leaders will be forced to think more deeply about how to make these new diversity team settings more effective than teams solely composed of human members. Companies that best manage the collaboration between humans and algorithms will be the ones that will be most successful in the future. And, be warned, it is not simply being able to promote the individual performance of the human member and the individual algorithm. No, it will depend on how well organizations can put the performance of both humans and algorithms together to create a mix that delivers something new.
This requirement makes clear that leaders in the future will be more responsible than ever for the success of their organization. They need to be able to tap into both their business and technological knowledge and use it to bring people and technology together. They will have to be intelligent in promoting co-creation between humans and algorithms. Mastering the ability of empowerment, as outlined in chapter seven, will be crucial. The ideal view on what I call empowering the new diversity is to ensure that both humans and algorithms achieve a collaborative relationship. The act of co-operation is understood as an effort in which members of a team, organization or society work together to achieve a common goal. It is a known fact that the human species survives not because of fighting with each other, but because of co-operation.
Co-operating in the new diversity
The same is likely to be true in today’s era of automation. Co-operation, rather than competition, between humans and machines will bring the most benefits to all. A nice example of successful co-operation between humans and technology is demonstrated in the case of Hyundai Motors.171 Hyundai is a Korean car manufacturer who employ many line workers. The work is tough and requires significant physical effort, which can lead to injuries and both physical and mental exhaustion. One of the strains of the job is that these workers have to do a lot of heavy lifting. Adopting the perspective that humans and machines can co-create, Hyundai has created exoskeletons – wearable robotics – that the line workers can use. The robotics developed allow workers to be more productive, while at the same time doing a better job in safeguarding their health.
What successful co-operation between humans and machines does is augment the abilities of individuals and teams. We talk about augmenting human abilities because the potential level of achievement due to the new diversity is higher than can be expected in teams composed only of humans. Another case that demonstrates how technology can augment human productivity can be found in the online company Stitch Fix in the US. Their algorithm-driven business model helps human employees to provide personalized styling service to customers. Data (preferences, body measures and budget) are provided by customers and analyzed by algorithms, which then provide recommendations. If a customer follows up these recommendations and adds further preferences to their profiles, algorithms make suggestions for future purchases. Using algorithms in this way makes for more effective salespeople and happier, more satisfied customers.
Within team settings, algorithms can also augment the processes that create output. Indeed, intelligent systems can be used to augment decision making which, in turn, can lead to better actions. An example which illustrates this is algorithms being used to help doctors make diagnoses, a practice that enabled doctors to deliver higher-quality individualized treatments. Algorithms can access individual profiles, analyze that data and find patterns across patient populations. This refined knowledge, combined with the human skills of intuition, empathy, and creativity, will make for better outcomes. For example, when examining how to increase cancer detection in the images of lymph node cells, it was shown that an algorithm-only approach had a 7.5% error rate, while a human-only approach had a 3.5% error rate. The combined approach, however, revealed an error rate of only 0.5%.172
To be an orchestral conductor
Like any team effort, co-operation between parties needs to be properly co-ordinated. And, this is where leaders in the future will earn their rewards. In the 21st century, one of the most important leadership abilities will be to act as an orchestral conductor, to have humans and algorithms working together in symphony. Acting as an orchestral conductor could even be taken literally, as demonstrated by the Chinese telecom giant, Huawei, who in 2018 explored whether this new diversity style of collaboration could employ AI-technology to complete Schubert’s Symphony No.8, originally started in 1822.173 The first two movements of the symphony were completed by the master himself, but the last two, for unknown reasons, were never completed.
To undertake the task, the company utilized the AI technology of the Huawei Mate 20 Pro smartphone. After having the technology study 90 compositions written by Schubert, Huawei’s AI technology translated them into code and worked to extend them. The AI also listened to the first two movements of Schubert’s Symphony No. 8 to analyze the key musical elements. The technology then combined its analysis with the knowledge acquired from Schubert’s pieces to end up with a new melody for the third and fourth movement.
The one-million-dollar question, of course, is whether the AI was able to decide which new combination best represented the emotions and atmosphere that Schubert wanted his music pieces to communicate. The obvious problem in this project is that AI cannot feel, nor is it possible for machines to understand the soul of an artist. The new diversity idea implies that at this point humans have to enter the work equation. The human employee asked to lead this specific project was the composer Lucas Cantor – who writes for DreamWorks Animation and is known for many movie music pieces. His primary task was to avoid the awful fate that the ending to Schubert’s Symphony No. 8 would sound like elevator music.
So, what did we learn from this experiment? The most interesting finding was that Cantor himself was very positive about the collaboration and compared it to actually working with another composer. So, the work experience from the human side seemed to be acceptable, even enjoyable. According to Cantor, one reason for this was because the AI composer had no ego, something that he would usually have had to consider when making changes to a human composer’s work. The AI-driven composer was never in a bad mood and never protested when Cantor sent back pieces with a request to re-work. Also important was the fact that, because of the work process and the incredible pace of AI itself, the 18-minute piece was written in just a few exchanges. Overall, efficiency went up, the work process seemed enjoyable and the end product was rated as good by a critical human audience.
This example helps us to understand what a leader will have to do when empowering the new diversity to drive better performance. As Venema recently noted: “One weakness remains apparent: algorithms are poor at interacting with humans in scenarios where subtle forms of co-operation are required.”174 Indeed, algorithms learn and generate knowledge by relying on observations and identifying data patterns, but do not think at deeper levels of intelligence, where emotions and empathy are at work. For this reason, guidance by humans is key when making decisions that involve common sense and awareness of the strategic, moral and social dynamics that are at play in organizational decision-making. As Bernie Meyerson, IBM’s chief innovation officer, said: “Humans bring common sense to the work; by its definition, common sense is not a fact-based undertaking. It is a judgment call.” 175
Initiating co-operation
In light of the reality that algorithms innately lack common sense, it is crucial that they are used in the right way. And this way has to be decided by the team leader. At the same time, humans, with their own unique capabilities, will have to be used in ways that are best matched with their profile and skill set. A first task for a leader is therefore one that involves task allocation. Leaders need to shape team structures by deciding which tasks are allocated to algorithms and which ones to humans. But this is not where it stops. One benefit of teamwork is that different ideas can be collected and provide more integrative solutions to problems.
As the Huawei example shows, a challenge for leaders will be to motivate human employees to make use of the information that algorithms provide to the team. As such, leaders will initially have to motivate human team members to accept algorithms as co-workers. Then, human employees need to go beyond acceptance to perceive and value algorithms as useful team members. Such a changed perception can only be achieved if leaders empower human employees to initiate interactions with algorithms.
Algorithms do not possess the qualities to initiate co-operative relationships, however, they can be included as a more active member once relationships have been established. In order to achieve integrative solutions, humans must be motivated to take the first step. Leaders therefore will have to empower human employees to take a more active role in their relationships with algorithms at work. The best way to inspire such active commitment is to ensure first of all that employees are knowledgeable enough about the intelligent technology with which they are supposed to work. This will require leaders to make their employees more tech-savvy, which, as we have seen in chapter seven, is labelled under continuous education.
Understanding the need for algorithms can empower humans to be more willing to use the input generated by those same algorithms. Therefore, in addition to the leader’s responsibility to make human employees accept algorithms as team members, they will also have to structure the work setting in such a way that the output generated by both algorithms and human agents is collected and made transparent to all. This requires leadership to aggregate the output of humans and algorithms into a collective decision. To achieve this, leaders will have to act as a kind of transactive memory, in which they serve as the binding glue between the different agents in the team. From this position, leaders must make all knowledge available, foster integration as supervised by human employees and ultimately decide on its use to achieve the goals of the team and company.176
Creating added value
All these requirements for the leaders of tomorrow center around the skill of promoting co-operation between humans and machines. As all our examples demonstrate, such outcomes can only be achieved if the value that is created by algorithms can be translated to the human context, where it can create value relevant to human end-users. As Dewhurst and Willmot (2014, p. 2) argue, “the contextualization of small-scale, machine-made decisions is likely to become an important component of tomorrow’s leadership tool kit.”177
One specific area where this co-operative ability can impact almost immediately is in the working of HR departments, which are important for recruiting the right kind of employees. Algorithms can provide many benefits in this process; they can be used to analyze the massive amounts of data collected over years to identify suitable candidates, or even to focus on the right kind of information in a recruitment session. For example, Jobaline, a job-placement site, uses intelligent voice-analysis algorithms to evaluate job applicants. The algorithm assesses paralinguistic elements of speech, such as tone and inflection, predicts which emotions a specific voice will elicit, and identifies the type of work at which an applicant will likely excel.
Finally, once the right kind of people are working in the organization, you want to keep them. A study from the National Bureau of Economic Research demonstrated that low-skill, service sector workers (where retention rates are low) stayed in the job 15% longer when an algorithm was used to judge their employability.178
Once the appropriate work setting is created, co-operation will be facilitated, but another important question remains. What to do once co-operation is achieved? Traditional business models assume that team outcomes will directly contribute to the goals and vision of the company. Leaders therefore need to know the key priorities to pursue and how to ensure that integrative solutions directly contribute to the pursuit of company goals. This will have to be achieved by communicating clearly to employees the reasons why they have to perform certain tasks alongside the ultimate goal.
At the same time, leaders need to be able to communicate with data scientists to ensure that the right priorities are used when calculation principles are coded for algorithms. These responsibilities underscore the importance for leaders to be able to connect with employees with different professional backgrounds, ensure that those employees use algorithms in efficient ways, and that value is created by co-creation between humans and algorithms.
A final question to address is what happens to the leader when priorities are known and co-operation is facilitated? If leaders have installed these processes, do they remain participative?
Indeed, if this point is arrived at, the leader should participate less and devote more attention to strategic thinking and development. A famous saying is that leadership is about making others perform better, so that they have to do less. When I mention this quote in my classes, many students’ express interest, but not necessarily for the right reasons. They are mainly attracted to the prospect of doing less!
This kind of interpretation always forces me to make clear that doing less in this context does not mean the same as doing nothing. On the contrary, it means that once your team knows what to do, and they have been empowered in the best ways possible, it is time for leaders to step back and devote their time to the more traditional tasks of leadership. This includes planning and developing the important strategic steps a company needs to take in the future to ensure its competitiveness and long-term sustainability.
170 Bersin, J. (2016). ‘New research shows why focus on teams, not just leaders, is key to business performance.’ Forbes, March 3. retrieved: https://www.forbes.com/sites/joshbersin/2016/03/03/why-a-focus-on-teams-not-just-leaders-is-the-secret-to-business-performance/#26ead3bb24d5
171 Owana, N. (2018). ‘Hyundi exoskeleton aims to cut workers’ strains, will be tested in factories.’ Retrieved from: https://techxplore.com/news/2018-10-hyundai-exoskeleton-aims-workers-strains.html
172 Wang, D., Khosla, A., Gargeya, R., Irshad, H., & Beck, A.H. (2016). ‘Deep learning for identifying metastatic breast cancer.’ Copy at http://j.mp/2o6FejM
173 De Cremer, D. (2019). ‘On the symphony of AI and humans in the work context.’ The World Financial Review, September-October, 61-64.
174 Venema, L. (2018). ‘Algorithm talk to me.’ Nature Human Behavior, 2(3), 173-173.
175 Captain, S. (2017). ‘Can IBM’s Watson Do It All?’ Fast Company. October 10. Retrieved from: https://www.fastcompany.com/3065339/can-ibms-watson-do-it-all
176 Hinsz, V.B., Tindale, R.S., & Vollrath, D.A. (1997). ‘The emerging conceptualization of groups as information processors.’ Psychological Bulletin, 121(1), 43-64.
177 Dewhurst, M., & Willmott, P. (2014). ‘Manager and machine: The new leadership equation.’ McKinsey Quarterly, 1-8.
178 Hoffman, M., Kahn, L.B., & Li, D. (2017). ‘Discretion in hiring.’ NBER Working Paper No. 21709. https://www.nber.org/papers/w21709?sy=709