A new report by McKinsey Global Institute, the private sector think tank affiliated with the global management consulting firm, suggests that automation could put as much as a third of U.S. employees out of work by 2030.

The report estimates that between 16 million to 54 million U.S. workers, and between 400 and 800 million workers worldwide, might be in need of a new job within the next 12 to 13 years, depending on the rate of technology adoption. A.I. might revolutionize the workforce at an even faster pace than prior periods of change, such as when employment shifted from farms to factories, and later from manufacturing to service-based economies.

Some professions are more prone to automation that others. Automation could easily take over manual labor, particularly in predictable environments, such as fast food preparation and machinery operation. The researchers found that 60% of all jobs could have at least a 30% threshold of tasks that could be automated. Automation will have less of an impact of jobs that require expertise (generally well-compensated, white-collar positions) or that depend heavily on social interactions. It may also have less of an impact in less developed areas, such as India, where human labor is cheap and tech adoption remains prohibitively expensive for the time being.

Some worry that automation could exacerbate already extreme levels of income inequality in the U.S. The report notes that if "reemployment is slow, frictional unemployment will likely rise in the short-term," which might exert downward pressure on wages.

McKinsey suggests that governments will need to account for the changes by offering comprehensive job retraining and more generous income supplements and subsidies. Unemployment insurance, public assistance, portable benefits, comprehensive minimum wage policies, universal basic income, and wage increases wedded to productivity are all options that governments could explore to mitigate the disruptive effects of automation, according to the report.

It's not all bad news, however. "There will be enough jobs for all of us in most scenarios," says one author of the report: new jobs may be created, companies may spend savings earned through automation, and an increasingly aging global population may increase demand for healthcare workers.

Other groups are also studying the potential impacts of A.I. on the global workforce. Some are more optimistic than McKinsey. A recent paper jointly authored by economists at Northwestern University, Stanford University, and College de France predicted that if A.I. becomes capable of generating original thought, it could result in rapid innovation, economic hypergrowth, and limitless GDP gains.

And an article published in November by the National Bureau of Economic Research explored why intensive research investment in A.I. has thus far yielded little change in productivity, suggesting that there are still broad gaps between A.I. developments and their implementation. The authors of this article compared the adoption of A.I. to the electric motor, which, though first produced in the 1880s, did not result in significant productivity gains until the 1920s. By this reckoning, the impact of A.I. may be more gradual than McKinsey expects.

And still other sources are simply monitoring the changes as they unfold. The "A.I. Index," an ongoing project created by researchers at Stanford University and M.I.T., for instance, tracks developments in A.I. with the goal of promoting understanding of their potential social impacts - and hopefully preparing for them.