May 21, 2024
Globally, People Expect AI to Increase Fake News and Information
Majorities Support Government Regulating
Use of AI
Written by: Jessica Theodule, Research Manager of Strategic Insights, Project Liberty Institute
Edited by: Jeb Bell, Head of Research and Strategic Insights, Project Liberty Institute
The latest report from the Project Liberty Institute finds substantial shares of people around the globe strongly convinced that artificial intelligence (AI) will exacerbate online misinformation and manipulation, even as it increases government surveillance in the real world.
However, people agree less on whether AI will ultimately exert a profound influence on how we live and work. Majorities in the developing and emerging economies are more inclined than those in advanced economies to believe AI will affect job availability and leisure time. Younger generations, too, are more convinced than their older counterparts that AI will transform society.
These insights emerge from a comprehensive survey of over 14,000 adults across seven economies, ranging from developing to advanced as defined by the International Monetary Fund (IMF): the United States, United Kingdom, France, Brazil, South Africa, India, and China.
Conducted between October 11 to 20 2023, the survey also finds substantial support for government regulation of AI, with majorities in all seven countries stating that it is important for governments to create laws and rules to govern the use of AI.
Many believe AI advances will impact jobs, fake information and surveillance
A majority of people in the countries surveyed expect AI to exert a major influence on the level of both cybercrime and online misinformation, even as most also believe AI will boost the use of facial recognition by governments to police public spaces and combat criminal behavior.
A median of 63% (excluding China, where the question was not asked) expect AI to lead to greater government use of facial recognition for both public safety and law enforcement. Three-quarters or more in Brazil (79%) and South Africa (77%) share this view, compared with more modest majorities in countries such as the U.S. (58%), U.K. (57%) and France (55%).
Despite widespread belief that AI will increase government surveillance, roughly six-in-ten across all seven countries voiced worries about AI fueling growth of both cybercrime and the spread of false information. For instance, 62% anticipate AI leading to more online scams and cybercrimes and around 61% believe AI will facilitate the spread of false information. Additionally, about 61% of respondents think AI will facilitate the spread of false information.
Expectations regarding AI’s impact on both the use and misuse of data vary considerably by country. More people in developing and emerging economies than in advanced economies believe AI will have a major impact on the amount of data collected by governments, as well as levels of online criminal behavior and fake news and information. In South Africa, for example, fully seven-in-ten think AI will have a major impact on the use of facial recognition (77%); the level of online scams (74%); and the spread of false information (71%).
Beyond an impact on the collection and use/misuse of data, more modest shares across the surveyed countries believe AI will impact the future of jobs, medicine and leisure time. A 53% median believe AI will have a major effect on the number of available jobs. This view is more prevalent in Brazil ( 63%), South Africa (68%) and India (58%), than in the other countries polled. Those in France (42%) and the U.K. (41%) are least convinced that AI will dramatically alter the labor market.
A median of roughly half (49%) believe AI will play a pivotal role in developing cures for cancer and other life-threatening illnesses. Brazilians (66%), South Africans (56%) and Indians (55%) are most inclined toward this view; citizens in China (47%), France (44%), the U.S. (36%) and U.K. (36%) are less convinced.
When it comes to AI’s effect on leisure time, only in Brazil do more than half (52%) say AI will make a major difference. The belief that AI will dramatically alter the amount of time people have for leisure activities is particularly limited in France (31%), the U.S. (30%) and the U.K. (25%).
Across the countries polled, younger adults, typically between 18-44 years of age are more convinced that AI will affect leisure time and job availability. A median of 41% of younger respondents anticipate a major impact on leisure time, compared with 35% of older respondents. When it comes to job availability, younger respondents are more than 4 percentage points more likely than older counterparts to say AI will have a major impact on the number of jobs available.
Mixed views on AI’s impact thus far
AI, in the form of recommendation algorithms and other predictive computer models, is already a factor of life for many people around the globe. Attitudes and narratives about AI – whether utopian or dystopian – are already forming. When asked whether AI has generally been good or bad for society, the prevailing view across the seven countries surveyed is that AI has exerted a positive influence on society (median of 46% “good for society”), compared with 14% who say the technology has had a negative social impact. Notably, 31% say AI’s impact was neither good nor bad for society.
Assessments of AI’s impact on society vary substantially by country. Majorities in China (78%), India (61%) and Brazil (60%) believe AI has been a positive force in their countries. This contrasts sharply with views in France (31% “good for society”), U.S. (27%), and the U.K. (26%), where a minority share of people say AI has had a positive social impact. South African views of AI, meanwhile, are lukewarm, with 46% saying AI has been a positive for society.
Within countries, assessments of AI’s social impact often vary by age. This is particularly true in the advanced economies polled, where younger adults tend to be more upbeat about AI than their older counterparts. In France, the U.S. and U.K., respectively, those 18-24 are 17 percentage points more likely than those 55-74 to describe AI’s effect on society as positive.
In South Africa, younger and older views of AI exhibit a 14 percentage point gap in the same direction. Meanwhile, in China age-based differences follow a similar pattern, but the gap in positive views of AI is only 6 percentage points.
In India (5 points) and Brazil (3 points), older adults are actually more positive about AI’s social impact than their younger counterparts.
Globally, strong support for AI regulation
Despite varying perceptions of AI’s impact, there is widespread support for government regulation of AI. A median of 88% across the seven countries surveyed say it’s important for governments to create rules and laws governing the use of AI, including 55% who deem this to be “very important.” Majorities say regulation is “very important” in South Africa (66%), U.K. (60%), China (57%), India (55%) and the U.S. (52%). Support for AI regulation is slightly weaker in Brazil and France (48% and 47% “very important,” respectively).
In most of the countries surveyed, older adults are more likely than their younger counterparts to say it is “very important” for the government to create rules and laws to govern the use of AI. Double-digit differences in the attitudes of older and younger adults are found in France (30 percentage points), Brazil (22 points), the U.S. (22 points) and the U.K. (21 points). Older Indians, too, are more supportive than younger Inidans of AI regulation, but only by 6 percentage points more.
By contrast, in South Africa and China there is a reverse pattern. In both countries, older respondents are less likely to feel that it is important for the government (or in China’s case, “society”) to create rules and laws to govern the use of AI. The age gaps range are similar in the two countries: 6 percentage points in South Africa and 5 points in China.
Methodology
Project Liberty partnered with polling firm J.L. Partners conducted an international survey of 14,220 adults aged 18-74 years from October 11-20, 2023. Participants were recruited by Cint from online survey panel providers in seven countries: Brazil, China (mainland), France, India, South Africa, the United Kingdom, and the United States of America.
The survey comprised a minimum of 2,000 respondents in each country, with samples designed to ensure national representativeness across age, gender, and geographic region. Weighting adjustments were applied to align the final sample composition more closely with each country’s most recent census data.
The samples in the U.S., U.K. and France can be taken as representative of their general adult population under the age of 75. The samples in Brazil, mainland China, India and South Africa are more urban, more educated, and/or more affluent than the general population. The survey results for these countries should be viewed as reflecting the views of the more “connected” segment of the population.
Note: Percentages in the text and tables may not total 100% due to rounding.