“AI will revolutionize our lives,” say many in the field, but the American public is not convinced. The findings of a recent Pew Research Center survey reflect a stark divide on the potential impact of artificial intelligence over the next 20 years between those working in the field and the broader public. In fact, 56% of experts believe AI will have positive effects in the United States, and many Americans disagree, with only 17% eager for the benefits AI will be able to deliver.
This gap does not just have national implications, it also has personal gains. Among AI experts, 76% believe AI will benefit them personally, but only 24% of the public believes the same. Even more concerning is the fact that 43% of Americans believe that AI will hurt them clearly showing increasing mistrust of the technology. The US public has become more concerned over recent years whether hands are pulling strings and, especially, what they’re pulling them to (from misinformation and deepfakes to job displacement and bias), according to Colleen McClain, a senior researcher at Pew, who spoke to Ars Technica in an interview.
The ascendancy of generative A.I. tools like ChatGPT has thrust artificial intelligence into mainstream conversations, but that attention has not been matched by widespread enthusiasm. Instead, awareness has spread fear. The proportion of Americans who feel concern outweighs excitement about AI grew from 38% in December 2022 to 52% by mid-2023, according to Pew’s research. The public’s discomfort touches on several realms, among them privacy, work and human bonding.
There is heightened anxiety over jobs. Over the next 20 years, almost two-thirds of Americans think that AI will displace jobs but only 39% of experts think so. The two groups agree that some occupations, such as cashiers and journalists, are at high risk, but they disagree about others. For instance, whereas 61 per cent of experts expect that truck drivers will lose their jobs, only 33 per cent of the public is concerned about that. These differences reflect the different views on what AI can do and its industry-disrupting potential.
But there is much probably too much optimism about some of the underlying transformative potential of AI, especially in fields like health care and automation, experts say. In an extended interview published by Pew, one expert described a growing sense of jubilation over AI’s capacity to improve the accuracy of the diagnosis of different diseases and boost the development of medicines for treatments like breast cancer. Still, even among experts, enthusiasm is tempered by worries over privacy and misinformation.
The conversation is made more complicated by gender disparities. Male experts are much more optimistic about AI than female ones. Meanwhile, more than half of male experts surveyed (63 percent) say they believe AI will be good for the U.S.; just 36 percent of female experts are positive about it, according to Pew’s data. Women also feel more strongly about bias and data misuse and underrepresentation in AI design. One Asian expert emphasized the need for diversity when it comes to nationalities: We need more representation, not just in terms of gender, but also in terms of ethnicities, places where people come from.
Representation issues are not unique to gender. AI design also reflects the views of White adults better than those of Black or Hispanic people, according to both experts and members of the public. Only 27% of specialists say Black adults’ views are well represented, and just 25% say that for Hispanic adults. This raises ethical challenges regarding the extent to which AI systems are inclusive, and how their implementation can potentially worsen already existing inequalities.
The public and experts also converge on regulation. Both groups fear that government oversight of AI will be inadequate, not too strong. Approximately 60 percent of Americans and 56 percent of experts feel this way, hints at wider doubts on whether policymakers can keep up with rapid advances in technology.
Private companies aren’t immune from scrutiny. Although 55% of experts and 59% of the public express little or no confidence diagnosing and/or using AI responsibly, the level of distrust varies by sector. University-affiliated experts are especially skeptical, with 60 percent saying that they have little to no confidence in corporate efforts, versus 39 percent of those at work in private businesses.
The results of the Pew study shed light on a widening chasm between technologists and the public they seek to serve. As further development on AI continues, experts do not abandon their hope for AI as another tool that will boost the evolutionary path of humanity, while people are afraid to lose their right for privacy, jobs or prevent equalisation of classes in society. The onus is now on creators of AI to narrow the gap with humans at their best through collaborative partnerships that celebrate diversity, creativity, and risk-taking, unafraid of the notion of machine learning in fields like film and design. Get your AI to rework their drafts and combine its hottest new air into potential scenarios, but also guide its hands over more realistic GitHub membership forms it needs to know the rules to abide by them.
