While the overall impact of artificial intelligence on productivity is almost certainly going to be positive, the widespread use of AI together with an ever-increasing pace of adoption will likely cause some skills to become less necessary and others more.
Technological and managerial skills will be more in demand at the cost of physical and routine work. The knowledge and education divide will widen. Simply working hard will not be enough to lead an economically satisfying life. This will hit middle-aged workers hardest, as their opportunity to retrain decreases and the return on new investment in retraining may not seem economically attractive.
This is a critical challenge to the social fabric of this country and beyond. According to some estimates, 10 to 15 percent of current work will be open to automation and machine learning by 2030. This would mean about 15 million to 20 million workers will need to be redeployed. This number globally becomes 400 million to 800 million workers, a shift that requires significant policy and economic planning. The emerging markets could be especially impacted, as lower-cost labor becomes less of an economic advantage.
In the broadest sense, AI is using self-learning machines to perform tasks and activities previously done by humans, at higher speeds, better accuracy or lower cost, often all three. This increased efficiency is similar to other transformations in technology like when cars and computers were invented, when email replaced sending postal mail and when smartphones became ubiquitous. There is one crucial difference: AI machines can “learn,” rather than simply use rule-based programming. This difference significantly expands the scope of what machines can do and amplifies the impact on the workforce.
Since the 1900s, the American workweek has shortened by about 50 percent. With rising income levels and changes in work profile, it is possible that our workweeks will continue to become shorter. If we are able to navigate the skill transition successfully, the more creative, meaningful jobs will be done by humans and the more routine jobs by machines. This could lead to higher job satisfaction and a more meaningful career. For those who can make the transition, increased productivity will lead to greater income.
To illustrate this point, the productivity level of American workers has increased by more than four times since 1950, real GDP has grown by eight times, and GDP per capita has grown five times. Also, about 15 million net new jobs have been added since 1980. And this was the period of computers, automation, electronic communication and e-commerce. Similarly, according to estimates, $6.5 trillion in value will be created globally by 2035 because of productivity gains from AI, which will also create $9 trillion in enhanced consumer demand, for a total of about $15.5 trillion in AI economic impact.
While the aggregate direction looks rosy, to achieve this attractive end goal, policy and corporate leaders will need to address certain important issues:
- Create responsive and accessible retraining programs. The implications for future skill requirements in AI will most likely continue to evolve. We will need ongoing retraining and skill-upgrade programs from both government and corporations.
- Provide adequate financial support to workers. Funding will be important during the transition, when some groups will be less productive because of temporary breaks from the workforce. Skill-upgrade programs will also need financial support.
- Overhaul the education system. It is time to rethink the current system which presumes years of education followed by years of work. Perhaps education periods should be interspersed with work periods. This may be enabled by the same AI technologies that are creating the need for it. AI may enable better and more-customized online programs, improved matching of skill needs and training needs, and enhanced labor mobility because of remote working, or more likely a combination of all of these will be possible.
We need to start planning for these alternatives now.
Puneet Shivam is President at Avendus, a financial advisory service.