Can the United States Win the AI Race?February 27, 2019 - 8 minutes read
The acceleration of artificial intelligence (AI) has piqued the interest of many countries such as the United States, Russia, and China. Now, they’re all in a race to become the foremost authority in the technology.
Recently, the US Pentagon released an unclassified version of its AI strategy—but will these plans be enough to outpace the competition?
A Matter of National Security
The US has a long, winding history with AI that predates the 1970s. Back then, the Department of Defense (DoD) began funding a small group of AI developers and researchers to make a radical concept a reality: Give machines autonomous intelligence.
Since then, the government’s interest (and investment) in AI has ebbed and flowed. But the technology’s potential could not be ignored. And in the past few decades, more of it than ever has been realized. Re-catalyzing interest in the technology around the world, these advancements have many countries believing that AI is their path to a more prosperous future.
The Pentagon sees AI playing a central role in American national security. This is why its new AI strategy prioritizes rapid AI adoption for each US military branch. For the plan to work, the Pentagon will have to work closely with the American tech sector, particularly when it comes to cloud computing resources and sourcing algorithms.
Dana Deasy is the DoD’s chief information officer. He explains that Chinese and Russian investments in AI applications for their militaries helped shape the US AI strategy: “AI will not only increase the prosperity of the nation but enhance our national security. We must adopt AI to maintain our strategic position and prevail on future battlefields.”
The Heart of American AI
At the center of the DoD’s AI strategy is the Joint Artificial Intelligence Center (JAIC). Established in June 2018, the JAIC acts as the hub of the Pentagon’s efforts to support AI’s incorporation into all military branches. Essentially, any DoD AI projects valued more than $15 million will be vetted at the JAIC.
The center will also work on its own AI initiatives with the help of US tech titans. According to federal contracting records, several famous organizations are already on board with the idea. IBM, Google, Oracle, and SAP are among companies that have expressed interest in involvement with future DoD AI initiatives.
Previous joint ventures between the Pentagon and the tech industry did not go over well with the public. Last year, thousands of Google employees walked out of their offices to protest the company’s work and involvement with Project Maven, an initiative that utilized AI to detect vehicles and other objects in military drone footage.
The controversy ended with the company not renewing a contract to continue work on this endeavor. Shortly after, Google CEO Sundar Pichai released guidelines for use of the company’s AI. While they expressly forbid work on weapons, the guidelines did allow for contributions toward other military projects.
In 2017, the JAIC was first proposed by the DoD’s Defense Innovation Board. Eric Schmidt, previous chairman and CEO of Google, chairs the board. Brendan McCord, the previous head of machine learning at the Defense Innovation Unit (a DoD organization focused on helping the US military embrace emerging tech), was the primary architect of the JAIC’s structure. He’s also the primary writer of the DoD’s AI strategy.
Project Maven’s Popularity
Lt. Gen. John “Jack” Shanahan acts as the director of the JAIC. He led Project Maven and thinks many of the JAIC’s future projects will be collaborations with tech companies. “Commercial solutions are available for most of the problems we’ve discovered in the past and will discover in the future,” he explains. “That is where some of the world’s best talent resides right now.”
While Project Maven wasn’t exactly popular with Google employees and the general public, it was a hit with the US military and is currently being integrated into the JAIC. William Carter, the deputy director of technology policy at the Center for Strategic and International Studies, believes this was due to the return on investment: “One of the most remarkable things about Maven was that it was so cheap relative to the power of the system that was developed.”
During their most recent budget requests, both the Marine Corps and Air Force discussed their intentions to expand the use of Maven algorithms by using them on “multiple unmanned aerial vehicles” to identify targets based on drone data. Each drone would be equipped with cameras that can monitor 40 square miles of area simultaneously.
Plans Set In Motion
While the JAIC’s future budget hasn’t been finalized yet, a recent Pentagon budget request document estimates the center’s budget for this year to be $89 million. In 2020, it’s anticipated to be $414 billion. These funds would not only go toward working with many tech companies but may also be used to set up satellite locations near the San Francisco development community.
As far as projects go, the JAIC already has several in the pipeline. One initiative is focused on training algorithms to predict maintenance needs for H-60 helicopters, a model utilized across various US military services. Another endeavor is training Maven-type algorithms to identify firebreaks in plane footage of wildfires so that disaster response efforts can be expedited.
Cybersecurity is also at the top of the Pentagon’s priorities and will probably garner some of the JAIC’s funding for new endeavors. In 2016, the DoD held a contest in Las Vegas that had bots trying to hack one another while fixing their own exposed flaws. The winner took home $2 million.
Although there’s clearly a lot of moving parts to these plans, ethics remains a top concern for all parties involved. Currently, the JAIC isn’t working on any autonomous weapons systems, but it may in the future. And any tech company’s involvement with this is sure to cause controversy similar to what we saw with Project Maven.
Rasha Abdul Rahim is an Arms Control and Artificial Intelligence Researcher at Amnesty International. She says that companies who work on these government contracts should be cautious: “Tech companies need to take steps to make sure they don’t cause or contribute to human rights abuses.”
What do you think of the Pentagon’s plans to expand on AI applications for the military? Do you think this will give America the edge it needs to be a leader in AI? Or will this only result in endless controversy? Let us know your thoughts in the comments!