AI can become a tireless co-pilot to intelligence analysts

Mary Legere
Retired Lieutenant General Mary A. Legere, former Senior Intelligence Officer for the U.S. Army, is a Managing Director for Accenture Global Defense, focused on bringing Accenture’s global capabilities, services and fed-fit best practices to national defense intelligence and cyber clients.


  • With the rapid development of technology, the intelligence community now has access to many new tools but it also faces entirely new threats and new defense challenges.
  • Cyber-attacks by malign actors are increasingly likely to be directed at civilian rather than military targets. To prevent such attacks, cooperation is required not only among allied countries, but also between non-government entities, the armed forces and the private sector.
  • Managing Director for Accenture Global Defense Mary A. Legere is a former Senior Intelligence Officer for the U.S. Army. She explains how the rapid development of artificial intelligence has been viewed in the intelligence community and how generative AI can help intelligence analysts.
You started your career in military intelligence more than 40 years ago, in the Cold War era. What has changed most since those days?

When I began my intelligence career in 1982, my colleagues and I were focused on understanding the military of the Soviet Union ground forces and their aligned states armies, as well as their proxies in the Middle East, Indo Pacific, and Central and South America. This work was difficult, critical but fairly stable, as it sought to reduce the risk of miscalculation between hostile nuclear states.

With the end of the Cold War, my focus expanded to include a new set of volatile security challenges, including threats to regional instability, failing states, sectarian violence, terrorist attacks, pandemics, and the increasing militarization of the cyber, information and space domains.

The one constant through all of this was the focus and constant evolution of technology enabling our adversaries. With the expansion of the internet, the rise of artificial intelligence, the explosion of commercial space, and advances in quantum technology, the pace and scope of this change increased exponentially, requiring a more focused effort by the intelligence community on the technologies that were disrupting our world.

How did this technological change affect your work as an intelligence officer?

Throughout my career I have always been conscious of the fact that everything I know about technology that connects us will change. We’re always trained to keep our eyes on evolving technology as it affects the discipline that we’re part of—I was a signals intelligence officer. So clearly the transition from radios to the use of internet over IP was significant. It changed everything about the way we were trained.

While enabling positive innovation in public health, communications, and other commercial sectors, what increasingly became apparent was the dark side of information tech with cyber-attacks, disinformation, invasive surveillance, and attacks against citizens, infrastructure, government, and non-government entities on the rise.

In the hands of malign actors or states, it became clear technology in the wrong hands presents catastrophic impacts.
For me, and for this current generation of intelligence leaders, it’s never been more important for our community and citizens to understand the power and potential of new technologies, and to enlist in collective so that nations can safeguard our citizens and national interests from harm.

What can the military do to defend civilians against cyber attacks?

The key to cyber is very much a whole-government approach. We have to work very well with state and local governments who have their networks that fall outside of the Department of Defense, and commercial entities that are often the easiest underbelly.

If you want to get off the fence, and target and steal intellectual property, you don’t come through the front door of the installation, you go to the contractor. So we’ve had to learn, and I think for most countries with modern cyber organizations there is a clear role for the military. But it is as a teammate when we’re talking about defending the cyber capabilities of our country and our citizens, and contributing to early warning.

It’s a serious issue. There are trillions of dollars’ worth of intellectual property stolen from the United States because of cyber vulnerability, which has a huge impact on our economy. It’s something we take seriously and it’s something that I know Finland takes seriously as well.

OpenAI’s release of ChatGPT on November 30 last year sent many tech companies scrambling to reassess their business models and plans. How did the intelligence community react to these key moments in the development of AI? Does new technology pose a threat to national security?

As a champion of expanding the use of AI to augment Intelligence, the US intelligence community has been leaning in on the expansion of AI by its agencies for many years. So when OpenAI ChatCPT arrived on the scene for general consumption this past year, the intelligence community had a healthy interest in how it might leverage it at some point, but also cautioned US intelligence agencies to proceed slowly until there is a full assessment of the risks, challenges, and a coordinated plan on how to set conditions for safe and responsible adoption.

As for whether this technology will pose a threat to national security, the prevailing assumption is yes.

Ungoverned or unchecked by legal constraints, there is little doubt that disruptive or malign actors will continue to use large language models to generate and flood the web with vast amounts of disinformation, attempt to poison public or proprietary LLS or the use of GPT for sophisticated targeting or social network exploitation. The intelligence community with its partners in government will be assessing those possible options and looking to track, mitigate, or create countermeasures.

Can you give us a few examples of how modern generative AI is used or could be used in gathering intelligence, or protecting it?

A potential use case that builds on and off of AI work that is already underway is the embrace of ChatGPT, when safe, to be a co-pilot to the intelligence community’s many analysts. As the analyst man-machine team emerges, the hope is that ChatGPT will be a co-pilot to help analysts with the first version of an intelligence assessment—using all data available to produce human-friendly reports that offer insights, at the speed of a computer. This would free human analysts to spend more time and space to apply their tacit knowledge and context to review the effort for correctness.

But before that, Chat GPT is going to have to learn that a black box [a system which can be viewed in terms of its inputs and outputs without any knowledge of its internal workings] to the intelligence community is a non-starter. We need to be able to understand what’s underneath the technology. But these issues are not intractable. We can send people to the moon. We can solve pandemics. We can solve these problems.

You’ve talked a lot about collaboration on cyber issues between the military, government, and the commercial sector and also between nations. This kind of cooperation doesn’t commence overnight on its own, I guess?

It started very rough, like all of us were. Like, do we really have to put a cyber force together? And the answer was yes. And it’s one of the remarkable things about Ukraine, that of all the many things they needed to get done in preparation for this inevitable war, they made the decision to create a cyber force. They had so many other things to do, but they knew that to get cooperation from other countries they had to create an organization with professionals. Nobody wants to hand cyber information to somebody who doesn’t know what to do with it. I don’t think any nation should get arrogant about cooperation with other countries.

You’ve talked about the importance of talent both during your military career and after that. How can governments and national armed forces compete for the best talent with Silicon Valley’s data giants?

First and foremost, we have very different missions, and it is hard to compete in general. It’s easier to recruit people into the military when there is a national crisis, like in your country where people want to serve. Right now, the propensity to serve among our young people is going down a bit. And so they want to do other things.

The generation that I joined was the Cold War generation. We were the sons and daughters of World War 2 and Korea veterans, so this was always something we thought of. The generation I commanded was the 9/11 generation, who left high school to join the military, and they’re now leading the military. This current generation has not had that existential threat. So in some cases they’re trying to figure out what their options are.

If you want to go and work in Silicon Valley because that’s your heart’s desire, you’ll do that. When you come into the military, what happens is once you’re in, you tend to stay because it’s the mission and the people that you’re working with. Some of the most creative, most brilliant people that are now working at some of the highest levels of information technology started out in the military, and they didn’t get out because they were allowed to do things for their country and that they would never be allowed to do on the outside.

We do have to make the point to work hard to get them in, and once they’re in, keep them challenged. You know, I always told myself I would get out if I wasn’t challenged. So there’s another really interesting set of initiatives right now exploring how we bring private industry into the military for short stem and also opportunities for people to leave the military to take a break. Maybe have children, look after aging parents, and then come back. We don’t have a lot of flexibility and we’re trying to do that now because we think it’s going to be important for non-traditional careers.

Finland has just joined NATO and signed a Defense Cooperation Agreement with the US. There’s been a lot of talk about military bases, F35s and nuclear arms, but less about what isn’t visible. What kind of possibilities does NATO membership offer regarding cyber defense and intelligence?

As part of its membership into NATO, Finland is now actively integrating into NATO’s intelligence and cyber enterprise, tying in their staffs and organizations with NATO’s organizations, and partner nations, as a full partner in that federation of nations that collaborate, produce and share intelligence and cyber intelligence and information relevant to the alliance, and to Finland’s specific security concerns.

While there will be fewer public pronouncements based on obvious concerns for operational security, those in NATO who lead the Intelligence and Security portfolios will be actively working with Finland Defense, Cyber and intelligence organizations to tie Finland to that enterprise – to ensure Finland can access, benefit and contribute to the work that NATO has been doing to, modernize and expand its capabilities in response to the dynamic threat environment that NATO must contend with and to improve the timeliness, relevance of the intelligence and cyber threat warning information.


Text by Tuomo Tamminen
Translation by Leni Vapaavuori and Nick Moon