Promising Start, but Few Details in House AI Report

Yesterday, the House Oversight and Government Reform Committee (OGR) Subcommittee on Information Technology released its long-anticipated report on artificial intelligence (AI). Despite its title (Rise of the Machines is also the title of a Terminator movie), the report avoids the hyperbole and fearmongering endemic to recent AI policy debates and offers a reasonably balanced and pragmatic assessment of the AI policy landscape. That said, it overlooks key areas where U.S. government leadership is needed.

The report, echoing CSIS’s March report, A National Machine Intelligence Strategy for the United States (MI Strategy), emphasizes the need for U.S. government leadership in AI, particularly in the face of major investments by China in AI technology and policy leadership. It also highlights some of the most important issues facing policymakers, including the challenge of maintaining U.S. competitiveness and managing risks like misuse and abuse, bias, and job displacement.

The main recommendations of the report reflect some of the key themes of the MI Strategy, calling for increased U.S. government funding for AI research and development (R&D), opening up U.S. government data sets, and engaging with stakeholders to develop strategies for education and workforce development. These recommendations are broad, however, and need to be translated into concrete, actionable policies.

Importantly, not all R&D investments are equal, and U.S. government R&D funding needs to be focused where it can add the most value. The private sector is already investing tens of billions of dollars each year in AI, but private investment is focused on short-horizon applied research in a small subset of AI fields like machine vision, natural language processing, and autonomous vehicles. U.S. government funding should focus on areas that are not adequately supported by private research funding and which have broader societal value, particularly long-term fundamental research. R&D funding can also be used to look for technical solutions to some of the AI challenges identified in the report, for example, by developing alternative approaches to AI that are more transparent and resilient and less data-intensive than today’s machine learning-based methods.

The report also reflects a key tension in the current AI policy debate between calls for stronger regulations to manage AI’s risks and concerns that over-regulation could throttle innovation in a critical emerging industry. It calls for federal agencies to take a more assertive approach to regulating AI’s risks under existing authorities, but it also acknowledges the risks of overregulation and suggests a measure of restraint in developing new regulations and authorities to govern AI.

This balanced approach is sensible, but the tension evident in the report hints at an underlying challenge for the AI industry and policymakers. While many in both communities prefer a hands-off regulatory approach based on private sector standards and best-practices, the tech sector’s credibility on managing issues of privacy, security, and broader social impacts of technology is at an all-time low. Without a substantial effort on the part of the major tech companies to develop credible governance and accountability mechanisms around AI and show that they can effectively manage AI’s risks themselves, regulators and political leaders may feel compelled to act.

While the report correctly highlights the importance of U.S. government investment in R&D and a thoughtful and measured approach to regulating AI’s risks, it fails to understand AI in the context of broader technological changes, and therefore misses key areas where U.S. government support is needed. The “AI Revolution” is the product of many major technological trends coming together: 5G networks providing new levels of connectivity, the internet of things (IoT) serving up unprecedented amounts of machine-readable data, including data on our physical environment, cloud computing offering a home for the massive data sets and computing power that drives machine learning applications, and next-generation robotics to translate the insights derived from AI into physical action.

Maintaining leadership in AI requires political leadership to spur 5G deployment, provide regulatory clarity around IoT devices and applications, incentivize broader cloud adoption by U.S. companies, and determine how to safely and responsibly integrate AI-powered robots into American lives and workplaces. The Chinese government has its own approach to gain leadership and is investing heavily in 5G deployment, IoT innovation, cloud adoption, and next-generation robotics, as well as a host of related technologies under initiatives like Made in China 2025.

The report also conspicuously omits the critical role of the U.S. government in international governance and diplomacy. Unlike other major disruptive technologies in the last century, AI development is not just happening in the United States. Many countries are developing policies around AI development and use, as well as broader technology policy issues with AI implications like privacy and data security. In the absence of U.S. leadership, other countries are adopting policies that are detrimental to U.S. interests and do not reflect our values, while many of our allies look to the United States for leadership but struggle to engage productively with the administration.

The report, coming from one of the leading voices in Congress on technology issues, Chairman Will Hurd, is a step forward, highlighting the importance of government leadership in AI and showing that congressional leaders are beginning to engage on this important issue. But a national strategy for AI remains elusive, and the report reflects the limited understanding of AI in the policy community and the challenge of grappling with an issue that is so complex and intertwined with charged political issues like privacy, education, jobs, inequality and the role of the United States in the world. With luck, others in Congress and the administration will take up this report’s call for leadership in AI and increase their engagement and understanding of these issues.

William A. Carter is a fellow and deputy director of the Technology Policy Program at the Center for Strategic and International Studies in Washington, D.C.

Commentary is produced by the Center for Strategic and International Studies (CSIS), a private, tax-exempt institution focusing on international public policy issues. Its research is nonpartisan and nonproprietary. CSIS does not take specific policy positions. Accordingly, all views, positions, and conclusions expressed in this publication should be understood to be solely those of the author(s).

© 2018 by the Center for Strategic and International Studies. All rights reserved.

William A. Carter