The CIA’s data-challenged AI imperative
Mastering the technology is key to being able to “disrupt those entities that do not,” said an agency cyber policy official.
The Central Intelligence Agency is developing its own Chat-GPT-like tool, but the agency is still struggling to manage its data and quickly adopt commercially available solutions, said Dan Richard, the CIA’s chief cyber policy advisor.
“One of the things that we are grappling with is data management. We assemble and review large amounts of data information and we are constantly looking for ways to be able to more effectively analyze, synthesize, and provide insights that we can get from that information out to the private sector,” Dan Richard, the CIA’s chief cyber policy advisor, said during a Billington Cybersecurity virtual event Thursday. “We are constantly on the outlook for better ways for us to manage our data, better ways for us to more efficiently and effectively use the data we have to get it to where it needs to more quickly and in a more efficient manner.”
That need to improve data management is grounded in how the CIA collects and integrates intelligence data other agencies use to inform policymakers and private companies.
“Although invisible to most, we actually support a lot of the information that [Department of Homeland Security], the Defense Department, [Director of National Intelligence], FBI are actually providing to the private sector and take that feedback and fuse that into all of the information that we possess,” he said.
The CIA is also struggling to rapidly adopt commercial tech solutions, like applications, instead of spending time developing them in house.
“Something that could take us several months to sort of assemble in terms of an application solution, the commercial sector has already taken advantage of it,” Richard said. “And what we're looking to do is leverage those solutions to more quickly address some of these problems that we're currently facing.”
The CIA isn’t the only agency struggling to manage its data. Earlier this year, the Director of National Intelligence issued a multi-year strategy that puts data and AI at the center to improve how the intelligence community achieves its missions.
“The intelligence community has been working on AI and artificial intelligence issues for over a decade. So this is an area that we have already been grappling with [on] how to take advantage of this technology and apply it for our mission,” he said.
The IC already uses AI to do things like scan the news. The CIA is building its own Chat-GPT-like tool to help with research and writing. Other intelligence agencies have talked up AI’s potential to improve their jobs, including analysis. And, the National Security Agency recently announced it’s opening an AI security center to focus on cyber threats, following a study to learn how the technology could apply to its missions.
But mastering the technology, and the data that fuels it, is critical, because it could give organizations and governments that do the upper hand.
“Those entities that augment their activities with AI applications will likely disrupt those entities that do not. So I think it's imperative that we find a way to tap into this technology to support the activities that we are entrusted with,” Richard said.
AI has clear “areas of concern” that “we have to address head on,” including cyber defense and disinformation, Richard said. “But those solutions cannot be at the expense of the innovation that we really need to more effectively and efficiently conduct the mission that we are asked to do.”