The White House has issued a comprehensive report addressing the potential legal and governance implications of artificial intelligence.
In October, the Office of Science and Technology Policy (OSTP) published a report addressing the challenges and opportunities in the developing field of artificial intelligence (AI). The report, Preparing for the Future of Artificial Intelligence, synthesizes feedback from industry experts as well as the public at large. As we recently discussed in our LawFlash Calling All Robot Enthusiasts: The White House Wants Your Input, the report is part of ongoing White House efforts to drive public dialogue on emerging issues in AI.
The report examines how the government, public, and private industry can work together to optimally advance AI technology—and implement it in a safe, ethical, and inclusive way. In so doing, the report surveys the AI landscape, discusses its existing and potential applications, and raises potential new policy concerns. Below are some of the key considerations highlighted in the report.
- “Freeing” big data. Many uses of AI for the public good rely on having sufficient data to train machine learning models and test AI system performance. In fact, the OSTP recommends that federal agencies prioritize open training data and open data standards in AI. Watch for a possible “open data for AI” initiative with the object of “releasing a significant number of government data sets to accelerate AI research and galvanize the use of open data standards and best practices across government, academia, and the private sector.”
- Leveraging AI for the public. Public- and private-sector investments in AI research and development have already spurred major benefits across fields such as healthcare, transportation, and the environment. For example, some cities have used AI to implement smarter traffic management infrastructure and increase urban mobility. Linking AI-enabled responsive dispatching and routing software—the same type used by ride-hailing services—with scheduling and tracking software has increased real-time access to public transportation and reduced wait times, energy use, and emissions by as much as 25% in certain cities.
- Regulating AI carefully . . . but flexibly. Federal agencies face an important challenge in regulating AI to safeguard the public, but without stifling innovation. To meet this challenge, the OSTP recommends that agencies have sufficient technical “seats at the table” in all regulatory policy discussions. Additionally, the OSTP suggests that policymakers adjust regulatory schemes to lower compliance costs.
- Welcoming new stakeholders. The OSTP recommends that the government invest in an advanced and automated air traffic management system that is “highly scalable, and can fully accommodate autonomous and piloted aircraft alike.” The OSTP also advocates for the US Department of Transportation to craft a nimble and evolving framework that integrates fully automated vehicles and unmanned aircraft systems (“drones”) into the transportation system.
- Optimizing collaboration with industry. Accurate and timely monitoring and forecasting of AI developments are crucial. The OSTP urges industry researchers to help the government identify milestones that could “represent or foreshadow significant leaps in AI capabilities.” At the same time, given that the private sector is not likely to invest in basic research and long-term, high-risk research initiatives, the OSTP recommends that the federal government prioritize investment in those areas.
- Diversifying AI. As the industry develops and expands, the OSTP highlights broad concerns regarding lack of diversity. Failing to integrate broad experiences, backgrounds, and opinions can lead to biased algorithms and unintended negative consequences. For instance, many caution against the use of perhaps biased algorithms in “risk prediction” tools used for sentencing and parole decisions. In other words, “the best AI is produced by diverse populations for diverse populations.”
With these considerations in mind, the OSTP hopes that a more informed and engaged public will make the United States truly “prepared for the future of AI.” To further this goal, the OSTP also plans to convene a study on automation and the economy to support a further report for release next year.