Artificial intelligence is transforming the craft of spying: it sharpens the ability to find, profile and groom human sources while giving adversaries powerful tools to create convincing fake identities, deepfakes and other deceptions that are harder to spot.
Jonna Mendez, a 27-year CIA veteran who once led the agency’s disguise program, told an audience at Infosecurity Europe 2026 in London that AI has complicated intelligence work in unprecedented ways. The same capabilities intelligence services use to sift vast amounts of information can be turned against governments and companies to target employees and insiders.
“Protecting yourself means finding the soft spots — where adversaries are coming in and how they’re convincing people to talk,” Mendez said. The underlying goal of intelligence work, she added, remains the same: persuade someone to cooperate and share information that cannot be obtained any other way. AI, she argued, simply makes it faster and cheaper to identify who to target and what will motivate them.
Mendez described how disguise and identity tradecraft evolved from physical masks and covert transformations to an era in which digital fakery and synthetic personas can be deployed at scale. During her career she oversaw highly convincing physical disguises — including animated masks that once fooled trained observers in the White House — and field techniques like “disguise on the run,” used in Cold War operations to change appearance in transit.
Today, she warned, deepfakes, social media and AI-driven profiling create fertile ground for online recruitment and manipulation. The Five Eyes alliance recently warned that Chinese military intelligence has been exploiting professional networking and job platforms by posing as recruiters, consultants or think-tank representatives to lure people with access to sensitive information. According to the joint bulletin, operatives rank applicants and target those likely to have access to classified or privileged material, focusing on defense, intelligence, academia, journalism and related fields.
Analysts also point to China’s massive camera networks and AI-powered facial recognition systems as another dimension of modern tradecraft: hundreds of millions of cameras linked with data streams can track individuals across cities in real time and flag anomalies.
Mendez stressed the continuing importance of trust in human intelligence. Tradecraft long relies on predictable motivations for betrayal — summarized by the acronym MICE: money, ideology, compromise, ego — and AI helps determine which trigger will work for a given person. But she emphasized that meaningful human intelligence cannot be done entirely remotely: “You can’t do it from an armchair. You’ve got to get up and cross borders.”
At the same conference, security practitioners described a concrete example of how adversaries are combining stolen identities, forgery and digital tools to penetrate organizations. Steve Povolny and Findlay Whitelaw of Exabeam recounted a case in mid-2025 where a North Korean operative posed as a young American IT professional, submitted doctored documents and passed hiring checks at a U.S. cybersecurity firm. Within 24 hours of being hired, Exabeam’s AI-driven user and entity behavior analytics system detected unusual behavior — installation of remote-access tools, connections to command-and-control servers and other red flags — and alerted analysts. The company shut down the activity and reimaged the laptop within hours, preventing data theft.
The Exabeam team said they went public with the incident to warn the industry, noting other firms reported similar intrusions after hearing the case. In 2025 the U.S. Department of Justice coordinated seizures and charges connected to operatives who used stolen and fake identities to obtain employment at more than 100 U.S. companies, and the FBI has warned that North Korean actors have extorted firms by holding stolen data hostage.
Mendez’s career provides a long view of how deception has changed. She joined the CIA’s Office of Technical Service in 1970, training assets to use miniature cameras and other covert tools and eventually leading a global disguise program aimed at Soviet and allied services. Her late husband, Tony Mendez, was the CIA operative behind the 1979 rescue of six diplomats from Tehran using a false film-production cover — an operation later dramatized in the film Argo.
Her message was not that technology makes human spies obsolete, but that AI multiplies both opportunity and risk. Agencies and companies must hunt for vulnerabilities in how people are recruited or persuaded, harden verification and hiring processes against synthetic identities, and combine human judgment with automated monitoring to detect anomalous behavior quickly.
In short, AI has elevated the spy wars: it amplifies traditional tradecraft while producing new modes of deception that demand updated defenses and vigilant, human-centered counterintelligence work.
