The Promise and Perils of Artificial Intelligence

This week the 47th Annual Meeting of the World Economic Forum (WEF) is taking place in Davos, Switzerland, bringing together business, government, individual activist, and non-profit leaders. The theme of this year's program is responsive and responsible leadership with a focus on five key areas: "strengthening global collaboration, restoring economic growth, reforming capitalism, addressing the identity crisis which has caused the current populist revolution, and preparing for the Fourth Industrial Revolution."

On the first day, a panel of leaders from business and academia assembled to discuss the impact of artificial intelligence (AI) and how businesses can design products and services that benefit society as a whole. The session was moderated by Robert F. Smith, Chairman and CEO of Vista Equity Partners, and the panelists were:

  • Ron Gutman - Founder and CEO of HealthTap
  • Joichi Ito - Director of the MIT Media Lab
  • Satya Nadella - CEO of Microsoft
  • Ginni Rometty - Chairman, President, and CEO of IBM

As one might expect, each panelist was excited about the potential impact of advances in artificial intelligence and highlighted specific areas where their organizations are making contributions. There was broad consensus in the belief that AI will change the world in very significant ways. In the swirling of optimism around the technology, each panelist also expressed a sober realization that there are significant challenges and risks that, if not managed properly, could contribute to a further erosion in job security and inequitable growth.

Technology Could Be The Next Target

In recent years popular sentiment towards globalization has become increasingly negative and many now blame this trend on the lack of economic development and social progress they are experiencing. As this frustration and discontent have grown a call for rejecting globalization and a rise in nationalist tendencies is poised to influence governments and organizations to curtail or reverse globalization efforts. There is a very real danger that, if not handled properly, the impact on the future of work coming from technological breakthroughs associated with the Fourth Industrial Revolution, including artificial intelligence, could be seen as the next target.

Estimates vary for how much of today's work could be automated by future breakthroughs, but few present a picture more rosy than at least two-thirds of physical and knowledge workers being impacted. A McKinsey article from July 2016 studied 800 occupations and estimated that 45% could be automated by already existing technology. Even if the macroeconomic impacts are positive in the long run, that will be little comfort for an individual who loses their job as a result of being made redundant by the introduction of a new technology.

How Can We Respond?

As was highlighted during the panel discussion there are design decisions to be made that will influence the future direction of the field. What purpose will direct research and development efforts? Which guiding principles will be incorporated? Which data sets will be used to train these systems?

During the first few decades of AI research, these were choices for computer science and academics to make. Going forward an inclusive approach should be taken to involve individuals from across society in order to gain from their perspectives and lessen unintended consequences.

A key design decision that each panelist seemed to embrace was a focus on augmentation versus replacement of humans. This may not technically be necessary for all applications of AI, but I believe it is the best approach for two reasons. One is that it reduces the likelihood of a societal backlash against the technology that would likely result from an increase in uncertainty and fear of being replaced by a machine. The second is that it is a step towards a future in which the creative potential and most human qualities inherent in each of us could be expressed more fully through a partnership with AI.

Lifelong Learning Is The New Norm

Although the augmentation approach of partnering humans with AI systems shows more promise there are series challenges associated with retraining for the skills of tomorrow. As with previous industrial revolutions, there is expected to be a temporal and skills gap. It may be many years before it is clear what new occupations will emerge. In the interim entire generations of people who may have been well qualified for the jobs of today will find themselves without the necessary skills. We simply cannot predict the occupations of the future, and it is becoming clear that the current educational and labor systems are not well suited for an increasingly dynamic and rapidly changing job market.

Multiple solutions have been proposed, but one thing seems clear. Lifelong learning will replace the current model of intense education early in life and minimal training thereafter. As valuable as skills will continue to be it is more important to teach people how to learn. They will need to use this skill continuously throughout their working lives to adapt to changing demands.

It is critical that education is provided broadly and equitably, and there is a growing realization that individuals may not be able to shoulder all of this burden. A partnership between individuals, corporations, and governments would better prepare us for the coming fusion of physical, digital, and biological worlds.

One mutually beneficial strategy may be to direct artificial intelligence at solving these problems. If personalized, affordable, and accessible educational systems could be developed by leveraging AI, such systemic solutions could be the answer to the very problems that AI's flourishing created. If a symbiotic partnership is formed, the challenges that have faced humanity through all previous industrial revolutions might be blunted. In collaboration with our AI creations, we can shape a future that is favorable both for individuals and society as a whole.

Closing Thoughts

The most meaningful benefit of artificial intelligence may lie in enabling humanity to be at its very best. The surplus energy and time could unleash a wave of creativity, unlike anything the world has ever seen. However, we must guard against the loss of identity, relatedness, competence, and security which many have found in their work. In designing the future, the innate psychological needs of humans must continue to be met, while educating people with the skills and understanding to adapt to new ways of working.

I am of the opinion that the benefits do outweigh the costs, but this is tempered by the belief that it is only because the benefits are so staggering relative to considerable costs. If the near-term impact is an increase in inequality and an erosion of job security, technologists and society alike may be unable to realize the full potential.

You can watch the full 57-minute video on the WEF site here.