Americans trust military and academics over technology companies on responsible AI development

Nearly half of Americans have some or a great deal of faith in the U.S. military’s capability to produce AI innovation that does not pose risks, while more than two-thirds have no confidence or not much in Facebook’s ability to do the same, according to the Center for the Governance of AI.

The UK think tank based at Oxford University’s Future of Humanity Institute found the public significantly more skeptical of the trustworthiness of technology companies than the military in its “Artificial Intelligence: American Attitudes and Trends” (PDF) report.

Perceptions of Microsoft are in line with those of the industry in general, with just over 40 percent saying they have “fair” or a “great deal” of confidence in their AI projects, according to the survey. Amazon and Google, meanwhile, are trusted slightly less than Microsoft, but more than Apple. Microsoft has been calling for government regulation of facial recognition, which may signal a sense of corporate responsibility to the public. Academic researchers are roughly as trusted as the military.

“There is no organization that is highly trusted to develop AI in the public interest, although some are trusted much more than others,” says Allan Dafoe, director of the Center for the Governance of AI according to Bloomberg.

The report also found low, though increasing, levels of public understanding of AI concepts among the American public, with the majority saying they do not believe Facebook photo tagging, Netflix and Amazon recommendations, and Google translate do not use AI. Overall, American’s tend to support AI development, but with careful management. Demographic characteristics account for significant differences in support for AI, with women and those with lower incomes more likely to oppose AI development, and people with computer science degrees more likely to support it, according to the report.

China recently created an AI ethics committee to develop guidelines for the technology’s development.

Article Topics

academic  |  artificial intelligence  |  biometrics  |  military  |  regulation  |  United States