The U.S. government wants to boost its artificial intelligence capabilities or risks being left behind by the private sector and China.
In the last two years, that’s meant new AI initiatives from the Pentagon, the Defense Advanced Research Projects Agency and the intelligence community. Now, the Intelligence Advanced Research Projects Activity is requesting information about research efforts on “cutting-edge machine learning techniques.” IARPA posted the formal request for information Dec. 4. The deadline for industry to submit information is Jan. 17.
“Of specific interest is the respondent’s knowledge of, and experience implementing, current, cutting-edge machine learning techniques,” the intelligence community’s research arm said. Respondents are required to have top secret clearances to work on the project, according to the IARPA listing.
In addition to its deep learning program, IARPA leaders want information about research into “future computing systems” that can self-learn. Such a move could have implications for improving government cybersecurity.
“The need for real-time (or near-real-time) analysis of massive amounts of heterogeneous data in this new era of explosive data growth has dramatically broadened the application space for advanced computers,” IARPA said. “The current volume and variety of data are already beginning to exceed the ability of today’s most advanced classical systems to deliver optimal solutions.”
Most cyber threat detection platforms use some form of artificial intelligence to create warning indicators, according to public and private sector officials. However, the U.S. government is behind the private sector when it comes to use of AI, said James Yeager, the public sector vice president at cybersecurity firm Crowdstrike.
“There is, by design, a more staggered type of approach to some of these advances in technology in the public sector, and as a result, the government is going to be behind the private sector,” Yeager said.
IARPA has a “very high-risk-but-high-reward approach to solving complex problems. They take a lot of time and take a lot of resources,” said Yeager. “But If they can come out of that research project with a silver bullet, it is going to benefit everyone.”
Andrew Laskow, a senior manager at Blue Prism, which provides AI products to federal government and defense agencies, said that in the U.S. government many people are “looking to AI for problems that they cannot solve.”
“There is still a misunderstanding at the highest levels of what AI can and cannot do,” Laskow said.
Public and private sector officials warn that AI-backed threat network indicators can overload users and create too many warnings.
Michael McGeehan, head of business development at Blue Prism, described intelligent automation being broken down into the “thinking side” and the “execution side.”
The artificial intelligence platform is the “thinking side” that makes decisions and is analogous to the human brain. On the other hand, robotic processing automation is the “execution side” that carries out tasks, like an arm or a leg.
Justin Lynch is the Associate Editor at Fifth Domain. He has written for the New Yorker, the Associated Press, Foreign Policy, the Atlantic, and others. Follow him on Twitter @just1nlynch.