Anthropic is now hiring specialists to track activists as global threats, raising fears that dissent is being labeled as terrorism
Broad categorizations.
The line between corporate security and mass surveillance seems to be blurring. According to Cyber News, Anthropic is now actively hiring for an enterprise intelligence specialist to track and investigate activism, and they’ve grouped this activity under the same “global threat” umbrella as terrorism and crime. This massive shift raises serious questions about how these companies view public dissent.
The role, which offers a salary between $180,000 and $230,000, explicitly tasks the new hire with identifying, assessing, and tracking activism as part of a broader security strategy, per The American Prospect. While the tech industry is no stranger to protecting high-value executives, this specific focus on monitoring activists marks a new level of integration between corporate intelligence and what some call a pre-crime approach.
In some cases, this would involve reporting suspects to the police before an actual crime occurs. This surveillance architecture was detailed in a podcast interview featuring Anthropic’s Global Security Operations Center Manager Keon Ellison, Security Operations Manager Zach Melvin, and James Neufeld, who is the CEO of Samdesk, a company Anthropic contracts for risk detection.
Anthropic says the idea is to predict incidents before they happen
During the conversation, Ellison shared an example where Samdesk provided intelligence that a protest near an executive’s location had been moved up by 60 minutes. Because of that heads-up, the company was able to reroute the executive to a service entrance, which Ellison described as a way to avoid what could have been a high-stress situation.
The company has been open about its desire to move from reactive measures to proactive, predictive threat management. In July 2026, Anthropic told The Wall Street Journal that it tracks concerning behavior over time, which allows them to catch escalation patterns early. This is a fascinating, but unsettling, development. While it’s easy to see why a company would want to protect its staff, especially after a man threatened to kill executives at the company’s lobby, the broad categorization of activism as a threat is where things get murky.
The stakes are high because the public’s relationship with AI is becoming increasingly strained. Recent data shows that 52% of Americans are more concerned than excited about the rapid expansion of AI in their daily lives. We’ve already seen incidents where the tension turned violent, such as when the home of OpenAI CEO Sam Altman was targeted with a Molotov cocktail in April 2026.
Prosecutors alleged that a 20-year-old from Texas was motivated by a hatred for AI technology. Similarly, there have been protests against infrastructure projects like the $500 million data center in Martindale-Brightwood, where the home of a local city council member was shot multiple times.
These events have led to law enforcement agencies privately warning about the emerging risk of anti-AI extremism
However, the definition of extremism is slippery. The job posting for the enterprise intelligence specialist has activism listed alongside terrorism and geopolitical instability. This suggests that peaceful protests could easily be swept up in the same security apparatus designed to stop violent actors.
The contradiction is even more apparent when you look at how these companies report users. In a recent case reported by The San Francisco Standard, Anthropic reported a user to the police after he told their AI model, Claude, that he had purchased an AR-15 and had the CEO in his sights. When the man was contacted, he claimed he was just “f—ing around.”
Interestingly, Anthropic refused to provide the police with the actual messages, citing internal policy. It’s a strange situation where the company uses its tools to flag potential threats but then keeps the evidence behind closed doors.
This isn’t just about protecting physical assets either. There is a growing push to designate AI infrastructure as “critical infrastructure,” putting it on the same level as water and power grids. If that happens, it would likely give these companies even more access to intelligence from federal agencies. Imagine a future where civic engagement against these data centers is framed not just as a local zoning dispute, but a threat to national security.
Even inside these labs, the atmosphere is tense
Former Anthropic researcher Jacob Coxon recently resigned, claiming the company is gambling with lives by developing superintelligence, while other researchers have admitted there is a non-zero chance that AI could kill all humans. When the people building the technology are expressing those kinds of fears, it’s not hard to see why the public is reacting with such intensity.
As the company tries to build its reputation as an ethical AI lab, it’s facing a difficult reality. It’s caught between its desire to be the “responsible” alternative to its competitors and the growing reality that its technology is sparking a massive, and sometimes desperate, public backlash.
Whether this surveillance approach will actually mitigate risk or just deepen the divide between the tech industry and the public remains to be seen. One thing is for sure, though: the era of the quiet, unchecked growth of AI is officially over.
Have a tip we should know? [email protected]
Follow The Mary Sue:
