White House tells intelligence agencies: Use more AI

In a first, US President Joseph R. Biden Jr. issued a national security memorandum today telling federal intelligence agencies they need to pilot and deploy artificial intelligence (AI) in an effort to boost the nation’s security.

The memo is directed at the National Security Agency, the Federal Bureau of Investigation, the Department of Defense, and the Department of Energy and specifically tells the agencies to use AI to track and counter adversaries’ development and use of the technology.

The edict also directs the agencies to ensure AI adoption “reflects democratic values and protects human rights, civil rights, civil liberties and privacy.” The plan is to coordinate efforts with US allies “to ensure the technology is developed and used in ways that adhere to international law while protecting human rights and fundamental freedoms.”

The memo is a very high-level mandate without a lot of needed guidance, according to Avivah Litan, a distinguished vice president analyst with Gartner Research. It’s also disappointing, she noted, that the Biden Administration and Congress have yet to come up with cohesive federal AI legislation that sets the rules and frameworks for risk assessments.

“For now, it’s a confusing maze of US Federal and state rules,” she said. “The agencies need to come up with governance councils, processes and a common technology stack that enforces use-case driven policies and ‘democratic values.’ It’s one thing to say, ‘Go use AI responsibly’ — it’s another to say what to use it for, how to use it, and which runtime guardrails and continuous out-of-band monitoring, evaluation, and assurances need to be in place.”

Jack Gold, principal analyst with tech industry research firm J. Gold Associates, believes the purpose of the White House memo is primarily to show the Biden Administration isn’t ignoring AI and is trying to not fall behind other nations, such as China and Russia — both of which are deploying it more for a variety of purposes.

Joel Meyer, president of public products at AI services provider Domino Data Lab, agreed, saying the memo is obviously aimed at preventing US adversaries from achieving “overmatch” by making the integration and application of AI in US military and national security capabilities an urgent priority.

“This memo takes an important step in both accelerating innovation and adoption and in ensuring that use is responsible and governed by putting in place guardrails for how US government agencies can, and just as importantly cannot, use AI,” Meyer said.

For example, Meyer said, the US Navy’s Project AMMO uses AI to support underwater target threat detection and to provide underwater drone operators with feedback data to increase operator confidence. “The [memo] builds a foundation of trust that allows programs like this one to scale by both accelerating innovation and adoption and ensuring that use is responsible and governed,” Meyer said.

The US agencies will also be responsible for increasing the security for, and diversity of, advanced computer chips to power compute-hungry AI models. The CHIPS Act is attempting to increase funding for new fabs and R&D facilities. To date, much of the money has been allocated, but not dispersed to chip makers.

“Our competitors want to upend US AI leadership and have employed economic and technological espionage in efforts to steal US technology,” the memo states. “This [order] makes collection on our competitors’ operations against our AI sector a top-tier intelligence priority, and directs relevant US Government entities to provide AI developers with the timely cybersecurity and counterintelligence information necessary to keep their inventions secure.”

The memo directs actions to improve the security and diversity of chip supply chains, and to ensure that, as the US supports the development of the next generation of government supercomputers and other emerging technology, the nation does so with AI in mind.

The government’s AI efforts will be overseen by the existing AI Safety Institute, which is housed within the National Institute of Standards and Technology (NIST) in Gaithersburg, MD. The White House said it’s the one body staffed by technical experts who understand the quickly evolving technology.

The order also lays out strengthened and streamlined mechanisms for the AI Safety Institute to partner with national security agencies, including the intelligence community, the Defense Department, and the Department of Energy.

The upside, regardless of whether the US takes a lead in AI development is that government and military are often leaders in experimenting with and deploying new tech, and in the process often kick-start a lot of technology innovation that gets used for normal business and consumer uses, according to Gold.

“That’s how the Internet came to be, so this is beneficial,” Gold said. “But AI not done right also has some potential negative results that could be harmful. It’s important to know not only what you are using, but how, to make sure it’s not leading you astray in your conclusions.”