"There will be less privacy, of course": How and Why People in 10 Countries Expect AI Will Affect Privacy in the Future

Abstract

The public has many concerns and fears regarding artificial intelligence (AI). Some are general or existential, while others are more specific with personal repercussions, like weakened human relationships, job loss, and further erosion of privacy. In this work, we provide a deeper understanding of how AI privacy concerns are taking shape. We surveyed public opinion of AI's expected effects on privacy with 10,011 respondents spanning ten countries and six continents. We identify four main themes regarding how the public believes AI impacts privacy: vulnerability of data, highly personal data and inference, lack of consent, and surveillance and government use. Unlike many aspects of AI and algorithmic literacy, for which public perception is often reported to be riddled with inconsistency and misconceptions, these privacy concerns are well-reasoned and broadly aligned with expert narratives. Based on our findings, we provide a roadmap of public priorities to help guide researchers and the broader community in exploring solutions that ameliorate AI's impact on privacy, and to inform efforts related to civic participation.