What happened
Anthropic has hired 33 employees for counterintelligence functions so far this year, up from 14 last year and nine in 2024, per Revelio Labs data compiled by Reuters; OpenAI's counterintelligence hiring rose to 22 in 2026 from nine in 2025.
Why it matters
With most counterintelligence salaries ranging from $200,000 to $500,000, these roles are becoming a costly standard line item for AI firms, and demand outstrips the supply of people with relevant experience, former Secret Service agent Christopher Falkenberg told Reuters.
WHO IT HITSSecurity and intelligence professionals leaving government for AI companies stand to gain most, with salaries of $200,000 to $500,000, while AI firms' HR and security teams face a thin talent pool that demand outstrips, according to Insite Risk Management's Christopher Falkenberg.
Summaries like this, every morning.
The hiring wave comes after a string of security scares. An OpenAI engineer was arrested for planting a listening device in a company conference room, and the Justice Department alleged he was communicating with a member of the Saudi royal family. Earlier, the scandal that brought down Jensen Huang as CEO of Nvidia involved his admitted romantic affair with a woman suspected of being a Chinese spy, though Nvidia's board does not think national security or company secrets were compromised.
AI firms have also been stacking their advisory ranks with former government officials. Anthropic appointed former CIA deputy director David Cohen and former National Security Agency cybersecurity director Dave Luber to its national security advisory council in 2025, while OpenAI named former NSA director Paul Nakasone to its board in 2024. The government side is moving too: the Pentagon is reportedly planning a spying program using "nonofficial cover officers" who pose as executives, and has built an elite agency for technology sabotage, though the Pentagon has denied those reports.
Former FBI counterintelligence operative Eric O'Neill told Fortune that protecting executives means teaching them to recognize manipulation, since approaches can start at a conference, over drinks, or on a dating app, and by the time a security professional raises concerns, the executive may already be emotionally invested and unwilling to listen.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
AT&T has nearly 500 agentic AI workflows in production, handles over 1 million transactions monthly, and proce…
Xilect, a robot-chip venture set up by GPU designer MetaX and humanoid robot maker UBTech Robotics, completed…
NVIDIA Corporation (NASDAQ:NVDA) placed 10th on a list of AI infrastructure stocks to watch after CoreWeave's…
The Motley Fool's analyst chose Astera Labs over Arm for 2027, citing revenue growth of approximately 115.2% t…
At a JPMorgan Chase investor event, Anthropic CEO Dario Amodei said compute is "the deepest business dilemma"…
Dell Technologies and the University of Alabama announced a high performance computing center for statewide re…
