NATO must recognize the potential of open-source intelligence

Air Marshal Sir Christopher Harper is a former UK military representative to NATO and served as director general of the NATO International Military Staff from 2013 to 2016. He is a nonresident senior fellow with the Transatlantic Security Initiative in the ’s Scowcroft Center for Strategy and Security and an adviser to companies, including Accenture and Adarga, which provide AI tools for processing open-source information, including for public-sector clients.

Robert Bassett Cross is a former British Army officer and the founder and CEO of the UK-headquartered AI software developer Adarga. He is a nonresident senior fellow at the Forward Defense practice of the ’s Scowcroft Center for Strategy and Security and an honorary research fellow at the University of Exeter’s Strategy and Security Institute.

Writing in 1946, just a few years before NATO was founded, Director of the US Office of Strategic Services Bill Donovan knew precisely how valuable publicly available information could be.

“[E]ven a regimented press,” he wrote, “will again and again betray the national interest to a painstaking observer . . . Pamphlets, periodicals, scientific journals are mines of intelligence.”

Today, seventy-five years after the Alliance was formed, such open-source intelligence (OSINT) is more important—and more powerful—than ever. However, underinvestment in OSINT capabilities and a culture favoring classified data currently hold back member states’ intelligence-collection potential. To fully utilize the available technology to detect threats from adversaries, NATO member states must overcome these barriers to embrace open-source intelligence enabled by artificial intelligence (AI).

Understanding the threat landscape

OSINT can help leaders get a fast, up-to-date understanding of their operating environment. If you want to know who’s doing what, where, and when, then an open-source specialist can quickly tell you.

If, for example, you want to find out who’s jamming GPS systems in the Baltic region, the relevant data isn’t hard to come by. Similarly, OSINT analysts can provide insights into issues ranging from the effectiveness of Iran’s attack on Israel (and the Israeli response) to China’s current role in fueling the Russian war machine. 

In recent years, it has become increasingly clear that, in addition to insight into current and recent events, OSINT can help leaders forecast what an adversary might be planning to do weeks, months, or even years from now.

By exploiting OSINT more fully and by integrating it into the wider intelligence cycle, NATO can preempt, deter, and defeat its adversaries’ efforts to expand their influence and undermine the security of member states. Here are several ways that OSINT can be used:

1. Across the physical domains of land, air, sea, and space, NATO can exploit publicly and commercially available data to explore an adversary’s order of battle and—more importantly—monitor changes in the strength and disposition of its military units and formations to infer its intent.
2. In the cyber domain, NATO can leverage commercially available information to detect and counter the penetration of networks governing critical infrastructure, as well as those related to research organizations, academic institutions, and technology developers.
3. In the information space, OSINT can help NATO identify, understand, and counter influence campaigns, specifically when it comes to the detection and attribution of disinformation and misinformation.
4. NATO can draw on vast swaths of open-source data to infer long-term strategic intent. Every subtle change to a government’s policies, every adjustment to its economic positioning and investment strategy, every new law and regulation it enacts, every new treaty and trade agreement—all of these can help the Alliance reverse engineer an adversary’s confidential playbooks.

Given the vast quantity, complexity, and diversity of the data, it is vital that NATO employs AI to extract the maximum value from it—to enhance analysts’ abilities, accelerate the analysis cycle, and build a reliable, contextual understanding of what Donovan called “the strategy developing silently behind the mask.”

The barriers to OSINT adoption

While AI is, of course, an emerging technology, its utility is already being realized across industries and sectors outside defense. From corporate intelligence and advisory services to finance and media, more and more private-sector organizations are using AI to make sense of the information environment, drawing on an ever-expanding range of sources to manage risk, identify opportunities, and adapt to geopolitical volatility.

However, the barriers to its widespread adoption and effective exploitation in political and military circles remain considerable. A paper published in 2022 by the Royal United Services Institute (RUSI), in collaboration with the Centre for Emerging Technology and Security and the Alan Turing Institute, identified three in particular.

First, there are tradecraft barriers relating to the methodologies governing everything from the analysis of publicly available information to the evaluation and dissemination of the resulting intelligence. Second, there are resourcing barriers stemming from underinvestment in the requisite tools, technologies, data sets, and training.

The third barrier identified by the RUSI authors—and the most daunting one—is cultural. Presented with so much open-source data, analysts and decision makers tend to favor classified information and internal data sets. These sources and insights are easier to trust and are imbued with what the authors call “the perceived power of the ‘secret’ label.” 

Speaking at the Eurosatory exhibition in Paris in June, US Major General Matthew Van Wagenen, deputy chief of staff for operations at NATO, confirmed how great this cultural barrier is. Up to 90 percent of “what Western militaries are looking for,” he said, can be derived from open sources:

This is a revolution in how we look at information. The ways of discerning information through classical means and techniques, tactics, and procedures that militaries have been adapted to—that’s really an old model of doing business. The new open source that’s out there right now, and the speed of information and relevance of information is coming, this is how things need to be looked at.

It is reasonable to believe that the tradecraft and resourcing barriers can be overcome. Methodologies are evolving swiftly, as are the requisite technologies. In fact, many of the tools NATO needs to capitalize on OSINT already exist. New AI applications are coming online almost every week. But if NATO fails to overcome the cultural barrier, it risks going into the next conflict underinformed and ill-prepared.

How AI-enabled OSINT can earn NATO leaders’ confidence

The cultural barrier to AI-enabled OSINT cannot be surmounted simply by decree or directive. Nor can it be overcome by intelligence professionals alone. The technology—and the discipline—must earn the justified confidence of civilian leaders and military commanders across the international staff, the military committee, and the supporting agencies. This could happen if AI-enabled OSINT were applied first to the simplest intelligence-gathering tasks before being applied to the most complex. To borrow the terminology made famous by former US Defense Secretary Donald Rumsfeld, NATO should apply the discipline to corroborating “known knowns,” resolving “known unknowns,” and surfacing “unknown unknowns.”

Corroborating “known knowns”: NATO should start by recognizing where the skills of the human analyst currently outperform even the most sophisticated models, and where AI can best be applied to elevate these skills. This means asking the right kind of questions, and employing OSINT to corroborate what is already known and to triangulate insights gathered from well-established secret sources. In this way, NATO can begin to overcome the skepticism that’s too often associated with publicly available information and OSINT. 

Resolving “known unknowns”: With so much data to draw on, it is essential that NATO uses AI to help collate, process, and (where necessary) translate that data so it is ready for analysts to interpret. If AI-enabled OSINT can prove useful to intelligence professionals in this capacity, those professionals may be more willing to apply it to the most complex and valuable intelligence tasks of all—surfacing risks and opportunities that civilian and military leaders would otherwise struggle to identify.

Surfacing “unknown unknowns”: Perhaps the greatest contribution that AI can make to the intelligence-gathering discipline is identifying patterns and connections that are invisible to the human eye. Dedicated, AI-powered information-intelligence applications that synthesize publicly available information with proprietary data can help analysts and decision makers tease out insights they would otherwise miss.

This combination of publicly available information with classified data will enable NATO analysts to give military and political leaders a uniquely rich, nuanced, and highly contextualized understanding of the operating environment. Decision makers at every level will be able to examine intelligence from every angle, and apply their experience and imagination to infer an adversary’s intentions based on the interplay of evidence.

The critical need for human-machine teaming

The necessary tools and methodologies exist. What’s missing is the determination to get these tools into users’ hands, to supply the requisite training, and to capitalize on the integrated output derived from all sources of intelligence, open-source and otherwise.

OSINT is becoming known among some intelligence professionals as “the intelligence of first resort.” Compared with clandestine methods of information gathering and analysis, OSINT is fast, low-cost, and low-risk. But if it can be combined with those same methods then NATO’s analysts and leadership will have an enduring competitive edge, with access to the kind of strategic information that would likely be, in Bill Donovan’s words, “of determining influence in modern war.”

NATO’s seventy-fifth anniversary is a milestone in a remarkable story of reinvention, adaptation, and unity. However, as the Alliance seeks to secure its future for the next seventy-five years, it faces the revanchism of old rivals, escalating strategic competition, and uncertainties over the future of the rules-based international order.

With partners and allies turning attention from celebrations to challenges, the ’s Transatlantic Security Initiative invited contributors to engage with the most pressing concerns ahead of the historic Washington summit and chart a path for the Alliance’s future. This series will feature seven essays focused on concrete issues that NATO must address at the Washington summit and five essays that examine longer-term challenges the Alliance must confront to ensure transatlantic security.

Image: Staff Sgt. Wendell Myler, a cyber warfare operations journeyman assigned to the 175th Cyberspace Operations Group of the Maryland Air National Guard monitors live cyber attacks on the operations floor of the 27th Cyberspace Squadron, known as the Hunter’s Den, at Warfield Air National Guard Base, Middle River, Md., June 3, 2017. (US Air Force photo by J.M. Eddins Jr.)