Research Roundup: May 2023
Could Machine Learning Increase Workloads?
Many people assume that artificial intelligence and machine learning will make their human workloads easier. However, a new study published in Management Science finds that machine learning could require human workers to exert greater cognitive effort than they did before the technology was introduced.
The study’s three co-authors include Tamer Boyaci and Francis de Véricourt, professors of management science at ESMT Berlin. They conducted the research with Caner Canyakmaz, formerly a postdoctoral fellow at ESMT and currently an assistant professor of operations management at Ozyegin University in Istanbul.
Using a machine-learning model, the researchers identified differences in accuracy, propensity, and levels of human cognitive effort required for both human-made and machine-aided decisions. The team found that, when used too liberally, machine learning can increase the likelihood of certain errors, such as false positives, in decision-making. This outcome would require humans to exert greater cognitive effort to reassess information and compare the machine’s conclusions to their own.
In healthcare, for example, machine learning is likely to improve the accuracy of doctors’ diagnoses overall. But if doctors rely too heavily on machine learning, the technology could increase the number of false-positive diagnoses, which could make it necessary for doctors to run additional tests and reassess the data.
The key, the researchers note, is for humans to use machine learning only as a complement to their own decision-making processes. “Human decision-makers are flexible and adaptive but constrained by their limited cognitive capacity,” which makes human cognition and machine learning great partners, says Boyaci. “However, humans must be wary of the circumstances of utilizing machines and understand when it is effective and when it is not.”
AI and the Need for ‘Relational Expertise’
As more companies adopt artificial intelligence and machine learning in their operations, many professionals in fields such as computer programming, education, and healthcare are focused not on whether AI and machine learning will make their jobs easier, but on whether these technologies will take over their jobs altogether. However, an article published in the Journal of Management Studies emphasizes that many professionals possess a skill that AI cannot replicate: relational expertise.
Relational expertise encompasses three areas. The first is the ability to take advantage of one’s ties and links with others. The second is the ability to know when to gather additional information about the problem at hand in order to develop solutions tailored to, and based on human interactions with, specific clients or patients.
The third is the recognition that the members of a profession are individually and collectively accountable for the outcomes of their decisions, advice, or treatments—whether automated or human-based. If the reasoning behind AI’s recommendations does not hold up, professional workers must intervene with an appropriate course of action.
The idea that machine-learning algorithms will eliminate the need for human professionals is “based on abstract and naïve notions of what these people do all day and how they do it.”
Workers whose jobs are affected by AI technologies should emphasize their skills in these three areas, advises lead author Pauli Pakarinen, postdoctoral researcher at Stanford University in California. Moreover, they should remain actively involved in the adoption and implementation of new technologies related to their work.
Expertise “is not mental or cognitive capacity that can be replicated abstractly,” says co-author Ruthanne Huising, professor and director of the Work, Technology, and Organization Research Center at emlyon business school in France.
The idea that machine-learning algorithms will eliminate the need for human professionals, she adds, is “based on abstract and naïve notions of what these people do all day and how they do it.” Pakarinen and Huising argue that roles that depend on relational expertise are likely to evolve with new technologies, not be replaced by them.
H&M Strategy Promotes Fair Living Wage
In a study published in the Journal of Accounting Research, scholars at the Rotman School of Management at the University of Toronto share the results of an innovative approach adopted by clothing retailer H&M Group. In 2013, the company was under increased pressure from activists to reform its global supply chain. As one response, H&M asked its suppliers to voluntarily adopt two programs.
The first program was designed to raise workers’ awareness of their rights and of formal opportunities to communicate with managers or labor representatives. The second was a wage management system that featured clear, transparent pay grids indicating fair compensation based on workers’ education, experience, skills, and performance. Concurrently, H&M implemented a system to track workers’ wages across 1,800 factories in nine countries.
After three years, H&M suppliers saw wages for workers increase by 5 percent, on average, or the equivalent of 44 USD per worker each year. That represents a significant return on H&M’s investment of 4.57 million USD—or 1.62 USD per worker—in the initiative.
The co-authors who analyzed the data provided by H&M include Jee-Eun Shin, an assistant professor of accounting, and Gregory Distelhorst, an associate professor in the Rotman School’s strategic management area. Distelhorst is cross-appointed to the university’s Centre for Industrial Relations and Human Resources.
Shin and Distelhorst found that suppliers who adopted both programs did not change their overtime policies or cut their workforces to pay for these increased wages. They did, however, receive an increased number of orders. The researchers speculate that the money to pay higher wages likely came from increased orders and productivity.
H&M could have taken the easy path and simply paid the workers more, but the company wanted to adopt a more sustainable, long-term approach, notes Shin. “Given the risk, there’s a lot of incentive for suppliers not to be part of this,” she says. Corporations have the power to improve conditions for workers worldwide, Shin emphasizes, so even the modest increase achieved through H&M’s entrepreneurial solution is “a huge step forward.”
Safer Spaces Lead to Bigger Startups
When looking for ways to grow their startups more quickly, many entrepreneurs focus on refining their marketing and growing their customer base. But a study published in the Journal of Business Venturing suggests that they should also spend some time creating psychologically safe spaces, where employees feel free to share their thoughts and empowered to craft their own roles within the company.
When employees feel safe to share their honest opinions with founders, about all aspects of the company, that feedback can lead to startups scaling more quickly. That is the conclusion of co-authors Evy Van Lancker and Mirjam Knockaert, both of Ghent University in Belgium; Veroniek Collewaert of Vlerick Business School and Katholieke Universiteit Leuven in Belgium; and Nicola Breugst of Technical University Munich.
To ensure their enterprises grow as quickly as possible, founders should ensure that employees feel safe and empowered to craft their own roles within the company.
The research team observed and interviewed founders and employees of an early-stage startup, in order to study how their roles developed over time. The researchers found that roles of founders and employees evolved over two phases—a “founder-driven phase” in which founders made all decisions about what to do themselves and what to delegate, and an “interaction-driven phase” in which the roles of the founders changed based on the types of interactions they had with their employees.
The researchers found that the more employees were able to influence their own roles and expectations as the startup transitioned from one phase to the other, the more likely the startup was to succeed. Moreover, founders fueled even greater growth when they also ensured that there was “a fit between their values and their employees’,” says Knockaert.
These findings indicate that, to ensure their enterprises grow as quickly as possible, founders should make employees feel safe and empowered to craft their own jobs, says Collewaert. In this way, founders will “get the most out of their own roles, but also the roles of their employees,” making startup success more likely.
■ U.S. teens are rethinking four-year degrees. The nonprofit Junior Achievement and banking institution Citizens have released their JA Teens and Personal Finance Survey, based on responses of 1,000 teens between the ages of 13 and 18. Among respondents, 76 percent agree that a two-year degree is sufficient to land a good job, with just 41 percent agreeing that a four-year degree is important for successful careers.
The biggest concern cited by respondents was the cost of postsecondary education—57 percent (compared to 49 percent in 2022). Their second most pressing concern was the need to take on student debt (50 percent). Teens were most interested in pursuing careers in healthcare (22 percent) or as digital influencers (18 percent). Less popular careers included those in government (9 percent), corporate offices (8 percent), and nonprofits (4 percent). More than 61 percent noted that they would rather start their own businesses than take on “traditional” jobs.
■ Professor develops AI-powered accountancy tool. Computer science professor Keyvan Shahrdar of Louisiana State University in Shreveport has developed a tool designed to ease accountants’ workloads. Powered by artificial intelligence, advanced natural language processing, and machine learning, ChatBotCPA can accomplish tasks such as reconciling accounts, tracking expenses, highlighting discrepancies between different data sets, filling in missing data on spreadsheets, and sending account updates to clients. By taking on such tasks, says Shahrdar, the chatbot “significantly reduces the effort involved and improves data accuracy.”
■ WRDS and SSRN name winner of research innovation award. Wharton Research Data Services (WRDS) and SSRN have named the College of Business at Prairie View A&M University in Texas as the 2023 WRDS-SSRN Innovation Award winner for the North America region. The school was honored “for its commitment to growth and innovation in academic research,” says Bob Zarazowski, managing director of WRDS. The award series recognizes rising business schools for their focus on impact-focused research and innovation.
■ Wits and Mastercard partner to study youth employment. The Mastercard Foundation and the Centre on African Philosophy and Social Investment (CAPSI) at Wits Business School in Johannesburg, South Africa, have partnered to study the outcomes of the foundation’s Young Africa Works strategy. The foundation created Young Africa Works with the goal of enabling 30 million young Africans to find fulfilling employment by 2030.
The Mastercard Foundation will provide 7.8 million USD to support collaborative scholarship that focuses on matching African youth with fulfilling employment opportunities.
The Mastercard Foundation will provide 7.8 million USD to the center over five years to support collaborative scholarship by researchers in 17 African countries. The project will measure the extent to which the nonprofit sector in these countries creates fulfilling employment opportunities for young people between the ages of 15 and 35.
Other objectives include training 20 doctoral students in relevant disciplines; generating knowledge, frameworks, and data related to youth employment; and creating a digital hub that matches young Africans with appropriate employment opportunities.
The foundation’s support, says CAPSI director Bhekinkosi Moyo, “provides us a once-off opportunity to test the hypothesis that the nonprofit sector is a major force to be reckoned with in African societies.”
Send press releases, links to studies, PDFs, or other relevant information regarding new and forthcoming research, grants, initiatives, and projects underway to AACSB Insights at [email protected].