David Patterson recently wrote a column in the Economist about how to face the threats of AI and how to materialize its capabilities. Anyway, according to Gallup, three-quarters of Americans believe that AI will reduce the total number of jobs within ten years. Patterson notes five principles to harness the AI capabilities in the best possible way:
1- When people and AI systems collaborate, they can do more than each could alone. AI applications aimed at increasing human productivity provide more advantages than those aimed at replacing humans. People who use more productive tools are more employable, satisfied, and have more opportunities. If the AI strays in areas for which it has sufficient training, humans can step in as safeguards. To put it briefly, concentrating on human productivity promotes the success of both humans and AI.
2-We should aim for productivity gains in industries that would generate more jobs in order to boost employment. In 2020, America has eight times as many commercial airline pilots and eleven times as many programmers as it had in 1970, despite massive increases in computers and passenger aviation productivity. Programming and air travel are two industries where demand is elastic, according to experts, which explains this increase. However, because agriculture is generally inelastic, productivity advances resulted in a three-fourths reduction in the number of farm employment within a human lifetime (1940 to 2020). Despite societal concerns, AI has the potential to boost employment if its proponents seek to enhance production in elastic industries.
3-The primary goal of AI systems should be to eliminate the monotony of existing duties. People will be more inclined to employ new AI technologies if they have more time for worthwhile tasks. Instead than doing a lot of paperwork, doctors and nurses select their professions because they want to serve people. Teachers would rather teach than grade assignments and maintain records. AI solutions that will increase the significance of people’s existing work in classrooms and hospitals should be given top emphasis.
4- Geographically, AI has different effects. While wealthy nations fear AI will replace highly educated workers, those with weaker economies struggle to find qualified specialists. AI may increase access to such knowledge in several areas, which might improve economic development and quality of life. For certain people in middle-income nations, improvements to vital services and local economies may even provide alternatives to leaving.
5- To assess AI advancements, we require improved measurements and techniques. The market can occasionally do this, like in the case of AI tools for expert programmers. This method cannot be used in high-stakes situations as we cannot accept the chance of hurting participants. A/B testing, randomized controlled trials, and natural experiments are the gold-standard methods that we must employ.
Source: Economist
