The rapid advancement of artificial intelligence (AI) systems has sparked a call for clearer guidance on governing the technology from public policymakers and industry leaders. According to a recent survey conducted by IEEE for the IEEE-USA AI Policy Committee, the majority of U.S. IEEE members believe that the current regulatory approach to managing AI systems is inadequate. They argue that prioritizing AI governance should be on par with other critical issues such as health care, education, immigration, and the environment.
The survey, which included a random sampling of 9,000 active IEEE-USA members and 888 active members working on AI and neural networks, revealed that there is no clear consensus on the definition of AI even among IEEE members. This lack of convergence on the definition of AI has significant implications for public policy.
One of the key areas of concern highlighted in the survey is the governance of AI-related public policies. The majority of respondents support protecting individual data privacy, regulating AI-generated misinformation, requiring risk assessments for AI products, and placing transparency requirements on AI systems. There is also strong support for restrictions on autonomous weapon systems, facial recognition, and algorithms in consequential decision-making.
When it comes to risk and responsibility, respondents believe that the public is inadequately informed about AI and that developers should bear the primary responsibility for ensuring the safety and effectiveness of AI systems. Trusted organizations for responsible design, development, and deployment of AI include academic institutions, nonprofits, and small and midsize technology companies.
In terms of comparative perspectives, there is a strong preference for regulating AI to mitigate social and ethical risks. While some professionals working in AI express concerns about regulation stifling innovation, the majority agree that it is crucial to start regulating AI immediately.
Overall, the survey results suggest a mixed governance approach is preferred, utilizing a combination of laws, regulations, technical standards, and industry guidelines to manage the development and deployment of AI systems. The findings will guide IEEE-USA in its advocacy work with Congress and the White House to ensure that AI is developed and managed responsibly to address societal concerns such as inequality and misinformation.