June 22, 2018
As companies experiment with artificial intelligence (AI), human-machine collaboration is emerging as a promising area of improved outcomes. While large companies tend to focus on the process automation side of AI, some have turned their attention to the interaction of humans and machines, says H. James Wilson, managing director of information technology and business research at Accenture Research.
In his podcast interview with AMA Edgewise, Wilson said that about 9% of companies adopting AI today are focused on human-machine collaboration—with significant results. “We find again and again in our research and work that teams of humans and machines really outperform pure automation,” said Wilson, co-author with Paul Daugherty of Human + Machine: Reimagining Work in the Age of AI (Harvard Business Review Press, 2018).
Wilson offers this example: A team of medical professionals created an AI-based method of detecting breast cancer cells. In this analysis, doctors were 96% accurate and machines were 92% accurate. But together, they identified 99.5% of cancerous biopsies in that health system, said Wilson.
For managers and employees, entering the world of AI can feel like they “are being parachuted into a dark forest,” Wilson said. They don’t have a trail map, and they don’t know what skills they’ll need to survive and thrive in that environment.
Wilson says that companies can design a number of AI-related roles to fill the “missing middle”—the area between humans and machines where collaboration can occur and new jobs are emerging. In this “wide, diverse, and job-rich area,” he said, humans and machines can improve each other and obtain exponential outcomes.
In its research, Accenture has seen that companies need to create six roles related to AI. The first three roles—trainers, explainers, and sustainers—build AI systems and manage them responsibly. The others—amplifiers, interactors, and embodiers—focus on augmenting employees through the use of artificial intelligence. These are “the mutually exclusive, collectively exhaustive set of AI roles that an enterprise needs if they want to effectively implement AI,” Wilson said.
In this environment, people may find that their job activities and workflow change. A scientist, for example, becomes an amplified scientist. “She’s now spending less time doing guesswork and routine lab work, and she’s able to spend more time assessing machine-generated insight and knowledge from experiments,” he said.
Listen to the podcast with H. James Wilson.
Visit the AMA Edgewise library for all podcasts.