The bots won't be coming for 007's job anytime soon. According to a former CIA officer, AI may help create false documents, but this fakery will give old-fashioned human intelligence fresh relevance.

Former CIA case officer (case officers recruit and handle foreign agents overseas) and RAND Corporation researcher Thomas Mulligan explained in the March edition of the CIA's Studies in Intelligence journal that, rather than render human intelligence work obsolete, AI may make human operators and analysts even more important to the intelligence community.

Humans won't just be necessary for intelligence work, Mulligan said. He lays out a case that AI will make human intelligence officers essential as truth becomes harder to distinguish from computer-generated fiction in the coming years. AI, Mulligan said, will be fantastic for counterintelligence operations, with intelligence gatherers, analysts, and case officers like him becoming ripe targets in the coming years. 

"Agents may already be using AI to fabricate information," Mulligan explained, adding that AI only makes the disinformation game easier to engage in. 

Sources could use an LLM to generate convincing false intelligence before meeting a gullible American handler, or an AI could even be used to dream up a convincing backstory for a counterintelligence agent seeking to plant false facts. 

"Widely available generative AI models put this capability into every fabricator's hands," Mulligan said. "Methods for detecting deepfakes exist and are improving, but so are the countermeasures."

Beyond that, he explained that AI has the potential to become the ultimate tool for the detection of intelligence operatives, with AI crawling camera feeds and being trained to spot every last little detail that could give away a spy. 

AI can have uses for the modern intelligence operative, Mulligan explained, with the tool being useful for identifying strategies for manipulating people, or for combing through mountains of intelligence for particular items. Case officers and other intelligence officials should be trained to use them, Mulligan argues, but also that AI will make the lives of field agents more difficult.

Welcome back to the age of spycraft

"It is tempting to think that, in an AI-saturated world, HUMINT [human intelligence] will be a relic," Mulligan posits in the paper. "The opposite is likely true." 

With AI-crafted misinformation threatening to overwhelm digital environments, AI undermining the trustworthiness of electronic communications, and technical intelligence cheap in the modern age, it's human-to-human spycraft tricks that are likely to become the standard for intel operatives, again. 

• CIA exposed to potential intelligence interception due to X's URL bug
• Chinese spies told Claude to break into about 30 critical orgs. Some attacks succeeded
• Attackers abuse Gemini AI to develop 'Thinking Robot' malware and data processing agent for spying purposes
• MI6, CIA using generative AI to combat tech-driven enemies

Instead of a call or an email, face-to-face meetings will be preferred. Dead drops of physical items are likely to remain in a spy's arsenal, as will brush passes of items in crowded areas.

In other words, AI is important in the world of spies, but it won't get rid of human spies. If that sounds like a familiar line of reasoning that people in the thick of an industry tend to say about their human employees, that's fair, but Mulligan told The Register that the world of spycraft isn't like other industries. 

"People believe it's just a matter of time before AI will be able to replace humans in the relevant work," Mulligan told us of other industries. "In the context I consider—HUMINT ops—there will be an enduring, robust role for human beings … in some ways the human element will become even more important in light of AI."

Keep an eye on those park benches, in other words: Even in the AI age, spycraft is still likely afoot. ®

More about

• AI
• Central Intelligence Agency

More like these