Insurers must think strategically about AI
Artificial Intelligence is dominating both headlines and the agendas of business leaders. Our 2018 Views from the C-Suite survey of global executives finds widespread agreement that there are tremendous opportunities in digitization and new technologies such as AI.
Fully 71 percent of executives expect AI to have "transformative effects for economic growth and competitiveness" over the next 12 months. However, executives may need to temper their expectations for the short-term implications of AI.
Much of executives’ enthusiasm is justified. AI is already being deployed in a range of arenas, from digital assistants and self-driving cars to predictive analytics software providing early detection of diseases or recommending consumer goods based on shopping habits. A recent Gartner study finds that AI will generate $1.2 trillion in business value in 2018—a striking 70 percent increase over last year. According to Gartner, the number could swell to close to $4 trillion by 2022.
This dramatic growth is likely reinforcing the perception among executives that such technologies can transform their respective industries. When looking at the external environment, encompassing economic, political, social and other external developments that affect business, one-third of executives flagged positive technological disruption in their industry as a top opportunity.
Among this subset of executives, 41 percent believe they are well positioned to respond or adapt to disruptive technologies, such as AI and big data. Survey results also show 42 percent of the same subset believes they can launch new business models—and 35 percent believe they can launch new products—that disrupt their sector.
Similarly, executives see successful adoption of new technologies and innovation in technology as profound opportunities to improve business performance. Big data, cloud computing, and AI/machine learning are the top three technologies that executives are looking to integrate into their operations.
Yet the C-suite also sees difficulties in adopting to the same new technologies in the short term—in our survey, in the year ahead. Slightly more executives view adoption of AI and other emerging technologies as a top operational challenge rather than an opportunity (38 percent versus 37 percent). This suggests a consensus among global executives that these technologies are important, but that it is not yet clear whether this will help or hurt their business in the short term.
These results also reflect an important reality about the current state of AI—it remains in its infancy. Deep learning holds great promise for exponential AI gains in the future but, as per Frank Chen, a partner at the venture capital firm Andreessen Horowitz, has observed, "we're in year two or three of a good, 40-year run … we have a long way to go just harnessing the existing techniques."
There is an array of evidence to support this idea. One the one hand, AI gains have been wide ranging in recent years, from advances in map applications and facial recognition, to AlphaGo Zero's skills in playing Go, and even the use of AI in creating original artwork. Yet there remains a long way to go in other areas, from Watson's lack of progress in curing cancer to continued flaws in voice assistants.
Progress has been made in "Artificial Narrow Intelligence," which performs a single task. Yet "Artificial General Intelligence" that exhibits human intelligence and "Artificial Super Intelligence" that surpasses all forms of human intellectual capabilities remain in the distant future. It will be crucial for executives to understand where AI is already impactful and where it is not quite ready before making their investment decisions. As they say, timing is everything.
Cybersecurity provides another example as to where AI is a double-edged sword. AI and machine learning are playing a role in protecting against cyber-attacks today, but they remain far from perfect—and they carry risks of their own. While use of such systems can help save time by automating cyber threat detection, they can also create a false sense of security among employees—especially when the AI algorithms themselves remain vulnerable to attack. Despite significant progress, AI is far from ready to fully replace human experts in cyber defense.
To be sure, AI continues to have a promising future. But executives would do well to set reasonable expectations for AI in the short term while preparing for an array of possible applications for the future. For now, this means focusing on limited AI applications rather than expecting human-like capabilities. Applying rule-based systems to solve specific problems rather than relying exclusively on opaque deep learning techniques can illuminate decision-making systems—and make course corrections far easier.
AI will not be a panacea for improved efficiency, competitiveness, and growth in the coming year. Nevertheless, executives are right to point to its transformative potential in the medium to long term.