American artificial intelligence strategy offers promising start

With artificial intelligence emerging as an issue in the 2020 presidential race, as Bernie Sanders thought it important enough to include in his campaign announcement and entrepreneur Andrew Yang has made the issue the focal point of his campaign, the White House announcement of an “American AI Initiative” this month could not have been more timely.

Recognizing that investment in artificial intelligence is “critical to creating the industries of the future,” President Trump issued an executive order directing departments and agencies of the executive branch to enhance American leadership in artificial intelligence with a coordinated federal government strategy. The initiative marks a significant first step toward a national artificial intelligence strategy, which should go a long way to address concerns that nations like China already have plans in motion.

{mosads}The most important component of the initiative instructs agencies to prioritize artificial intelligence in their research and development missions, encompassing not just publicly funded work in artificial intelligence, but also research and development within the federal government itself. That is significant because the federal government is inherently slow to adopt most new technology and is in need of the greater efficiency and innovation that artificial intelligence can provide. The initiative also directs agencies to prioritize educational funding and programs that help American workers and students learn relevant skills.

Finally, the initiative tells departments to establish guidance for safe and trustworthy artificial intelligence development and use, and to make their data sets and computing resources available to our American technology researchers and industries. This last point may not sound very sexy but, as a former scientist in the artificial intelligence field, I can testify to the vital importance of large data sets, which the government has a wealth of, and high speed computing resources to the success of artificial intelligence.

For those who believe in American exceptionalism, it is encouraging to see that the initiative unabashedly aims to ensure that the next great artificial intelligence inventions are “made in the United States” and calls for the creation of an action plan to protect the American advantage against “strategic competitors and foreign adversaries.” For those of us concerned with killing the golden artificial intelligence goose through stifling rules, such as algorithmic transparency, it is equally encouraging that the executive order imposes no regulatory regime. What the initiative does say on this is unobjectionable in directing agencies to develop both regulatory and nonregulatory approaches. One of the objectives is to reduce barriers, which presumably might include regulatory barriers.

Though discussion of potential risks dominates the public debate about artificial intelligence, the greatest of its dangers, such as out of control robots, are highly speculative and very unlikely to occur in our lifetimes. The initiative reflects that reality by taking an optimistic but balanced approach rather than focusing on these fears. It reminds us that artificial intelligence “promises to drive growth of the United States economy, enhance our economic and national security, and improve our quality of life,” but also includes that it should protect “civil liberties, privacy, and American values” and foster public trust among the guiding principles.

Similarly, the focus on educating American workers signals a proactive rather than fearful response to the concern that artificial intelligence will lead to the loss of jobs, a worry that is probably overblown but common enough to warrant inclusion in the Super Bowl ad for SimpliSafe earlier this month. The positive approach taken by the White House includes an emphasis on the promise of artificial intelligence to bring “great benefits for American workers, with the potential to improve safety and increase productivity.” Michael Kratsios, deputy assistant for technology policy to President Trump, also reminds us of the fact that artificial intelligence can “empower American workers by liberating them from mundane tasks.”

Critics of the initiative have focused on the lack of funding specifics. But the first step of a national artificial intelligence strategy must necessarily be written in general terms, and only Congress can appropriate money. Nonetheless, William Carter, a technology policy expert, wrote, “If they can find $5 billion for a border wall without Congress, surely the Trump administration can find a few billion to build the next engine of economic growth,” an argument that inadvertently points out the challenges of a president spending money without legislative approval from Congress.

Critics need not worry because we will soon have a lot more details about funding. The executive order requires that the agencies that perform or support artificial intelligence research and development to estimate the “total amount of such funds that will be spent on each such program.” The White House budget request for 2020 might also shed light on artificial intelligence funding. In the meantime, the initiative is off to a good start.

Curt Levey is president of the Committee for Justice. He previously worked in the artificial intelligence field and pioneered technology for overcoming the black box nature of neural networks that was hindering their adoption.