A new national survey commissioned by Northeastern University and conducted by Gallup finds most U.S. adults have an overall positive view of artificial intelligence, but believe they are ill-prepared to deal with AI’s expected impact on the global digital economy. The survey comes on the heels of numerous international studies forecasting significant job loss resulting from AI.
Overall, 22 percent of Americans with a bachelor’s degree or higher level of education say their college or university studies prepared them well or very well to work with AI. Moreover, only 18 percent are extremely confident they could secure the education needed to obtain a comparable job should they lose their current position to advances in new technology.
“The answer to greater artificial intelligence is greater human intelligence,” said Northeastern President Joseph E. Aoun. “The AI revolution is an opportunity for us to reimagine higher education—to transform both what and how we teach. If colleges and universities can adapt and modernize, we can ensure that tomorrow’s learners will be robot-proof.”
Americans optimistic about AI, expect negative impact on jobs
According to the survey, 79 percent of Americans say AI has had a “mostly positive” or “very positive” impact on their lives thus far, and 76 percent agree or strongly agree AI will fundamentally change the way people will work and live over the next 10 years.
Despite the optimism, a clear majority of Americans, 73 percent, expect the increased use of AI will eliminate more jobs than it creates. Additionally, 63 percent say the emergence of new technology and smart machines will widen the gap between rich and poor in the U.S.
Notably, in terms of their own careers, more than three-quarters (77 percent) are “not at all” or “not too” worried that their own job is at risk from advances in AI.
Americans favor employers over colleges for additional education
For additional education to secure a similar job to the one they lost, about half (49 percent) of all currently employed Americans say they would look to an employer for on-the-job training. Another 37 percent say they would consider turning to a college or university program, in-person (21 percent) or online (16 percent), respectively.
“With only two out of 10 U.S. adults thinking of turning to in-person programs at colleges and universities, there is both a need and an opportunity to reimagine all the ways ‘higher education’ can be done,” said Brandon Busteed, executive director of Education and Workforce Development at Gallup.
More than six in 10 Americans (61 percent) believe it is the responsibility of employers to pay for any retraining that results from automation eliminating jobs, while 50 percent say the burden of additional education should also fall on the federal government.
Blue-collar and white-collar workers differ
Differences in attitudes about AI between white-collar and blue-collar workers is the most pronounced of any subgroup, with blue-collar workers consistently more pessimistic about the impact of technology on people’s lives and work over the next decade.
Blue-collar workers are more likely to say AI adoption will result in net job loss than white-collar workers (82 percent to 71 percent). Twenty-six percent of blue-collar workers are worried about their own personal job loss due to automation, robots, or AI, while 19 percent of white-collar workers share that sentiment. Furthermore, just 34 percent of blue-collar workers are confident or extremely confident they could acquire the education or training they need if they were to lose their jobs, compared to more than half (51 percent) of white-collar workers.
Other noteworthy findings include:
Future of travel
- Self-driving cars: Calling into question the widespread adoption of self-driving vehicles, 42 percent say they would be “extremely uncomfortable” riding in a self-driving car. Moreover, 62 percent of respondents say they would be uncomfortable sharing the road with self-driving trucks.
Skills required to succeed
- Soft vs. hard skills: Nearly half of all Americans (49 percent) believe “soft” skills, such as communication, creativity, and critical thinking, are the most important for workers to possess to avoid losing their jobs to AI. Alternatively, 51 percent believe “hard” skills, including math, science, coding, and working with data, are the most important.
- Differences among age groups: When broken down by age, 62 percent of the millennial generation, Americans ages 18-35, believe “soft” skills should be a priority for workers. In contrast, 62 percent of Americans ages 66 and older say “hard” skills are the key to success.
Strong support for oversight
- Government regulation: A clear majority of Americans, 70 percent, agree or strongly agree that in order to protect consumers’ personal information and privacy, the federal government should do more to regulate companies that utilize AI.
Economic fears of immigration and offshoring vs. AI
- Source of job loss: Fifty-eight percent of respondents believe AI poses a more significant threat to job loss in the U.S. than immigration or offshoring (42 percent).
- Immigrants not a threat: Only 4 percent of Americans say they are very worried about losing their job to an immigrant. However, 37 percent say they personally know someone who lost their job to an immigrant or because the job was shipped overseas.
Universal Basic Income
- Slight opposition to UBI: A narrow majority of Americans (52 percent) disapprove of adoption of a universal basic income program to help support those who lose their job to new technology.
- Notable subgroup support: In contrast, majorities of those with less than a bachelor’s degree level of education (51 percent), millennials ages 18-35 (54 percent), and those who indicate they personally know someone who lost their job to technology, automation, robots, or AI (52 percent) support UBI.
- Make AI companies finance UBI: Of the 48 percent who support UBI, a large majority (80 percent) agree that companies that benefit the most from AI should pay higher taxes to support a UBI program.
The findings are based on a random sample of 3,297 U.S. adults ages 18 and older living across the United States surveyed by mail from Sept. 15 to Oct. 10, 2017.
The survey package included an English and Spanish survey to provide respondents the flexibility to reply in their preferred language. Gallup weighted the data to match national demographics of gender, age, education, race, Hispanic ethnicity, and region. In addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.