AI Predictive Crime Programs Take Shape As Public Opposition Mounts

As Americans turn against artificial intelligence, AI companies are building programs to surveil and identify troublemakers in advance.

An investigative report published this week by The American Prospect revealed that, “in addition to monitoring activists in the vicinity of Anthropic executives and keeping tabs on protests near physical Anthropic assets, the firm is also implementing a ‘pre-crime’ approach, attempting to predict incidents before they happen. In some cases, that also means reporting suspects to police before a crime occurs.”

An Anthropic security official said during a podcast that the goal is to transform operations “from reactive information to gathering proactive and predictive and preventative threat engagement and management.”

The Prospect garnered its findings by reading through job postings, talking to senior security officials at Anthropic, and listening to interviews with company officials.  

A Response to Growing Threats?

Anthropic frames its emerging surveillance and predictive programs as necessary responses to growing threats, and the Prospect report includes several examples of threats or concerning incidents involving Anthropic representatives. For example, last month, local media reported that Anthropic called the police on a man who told Claude that he bought an AR-15 and “had CEO Dario Amodei ‘in his sights.'”

Interestingly, while Anthropic was quick to call the cops, the company also refused to show police the messages. “In short, Anthropic reported a user for in-platform speech, and then refused to provide police with evidence of actual wrongdoing,” the Prospect pointed out.

In July, The Wall Street Journal published a report titled “The AI Backlash Has Tech Executives Fearing for Their Lives.” The Journal report starts out by relaying this incident:

A security guard at Anthropic rushed to stop the man sneaking into the lobby of the world’s most valuable AI startup. The man had entered by following closely behind a badge-swiping employee. He showed the guard an envelope marked with the name of a top Anthropic executive. The executive was “going to be killed,” he told the guard, and he needed to warn someone.

That incident, the Journal noted, “happened five days after an attempted firebombing of OpenAI Chief Executive Officer Sam Altman’s house.” That suspect has been charged with attempted murder and attempted arson. The police found a manifesto advocating the killing of Altman and investors.

Then there’s this. The Journal again:

In June, Anthropic security officials reported an Oklahoma man to police after he threatened violence while seeking a refund, according to police records. He wanted to talk to a human. “Since yall refuse to have a real person to contact me and refund my money ill be coming to your office with my pistol and then we will have a f—ing talk about my money,” the man wrote.

More Than Reacting to Possible Threats

So this is the justification for building these Minority Report-style programs. But Anthropic and others are using predictive technology for more than sniffing out lone firebombers and annoying AI protesters. Anthropic’s Global Safety, Intelligence, and Security (GSIS) team is also utilizing the programs. A job posting for an “enterprise intelligence specialist” for the team said Anthropic was looking for someone to “identify, assess, track, and investigate global threats including geopolitical instability, terrorism, crime, activism, nation-state targeting of the AI sector, and emerging security trends, including deep-dive research and [open-source intelligence] collection on specific threats, actors, and events.” (Emphasis added by The American Prospect.)

Anthropic has several contracts with the U.S. government, which has repeatedly proven to have no regard for individual privacy or, in many cases, even due process. So the implication with this kind of technology is hard to miss here. How long before the police start interrogating people who complain about AI, data centers, and surveillance cameras at the behest of a machine?

Nothing New

In response to the Prospect report, investigative journalist Whitney Webb, who has written some of the most in-depth reports on the merging of technology and power, reminded her social-media followers that predictive policing has been around for a long time:

Sorry, but it’s not just Anthropic. Palantir has been working on predictive policing and surveillance of dissidents for the US government for well over a decade. The goal of the Bush admin immediately after 9/11 was pre-crime via “Total Information Awareness” and they have been building it ever since, through the privatization of that program as Palantir. And there are other companies working toward this end too, all with ties to US intelligence or top US officials (Satellogic, SpaceX, Axon/Carbyne, etc.)

Anthropic may be “building” their own private equivalent, but this is something the government and its other private contractors have already built. Under the first Trump admin, Bill Barr as AG created the legal framework for pre-crime. They are just waiting on the right excuse or justification to scale it.

Problems With AI

The irony is that this isn’t likely to endear the public to AI companies. More likely is that it will inflame concern over AI and opposition to it.

On Thursday, Oren Cass, the chief economist at American Compass, summarized why Americans resent this technology and what AI companies can do to turn public sentiment around.

“Less than a quarter of Americans expect A.I. to have a positive impact on K-12 education, arts and entertainment, how people do their jobs, the economy, personal relationships or happiness over the next 20 years,” writes Cass. He adds:

For the past two decades, Silicon Valley’s leaders have treated their fellow citizens like caged guinea pigs, bombarding us with inappropriate and culturally corrosive content, turning a blind eye to social dysfunction and outright addiction, flooding schools with technology that hooked young users but did little for education and then sheltering their own children from the products that they marketed to ours.

And that’s just the beginning. “Chatbots frequently give wrong answers and terrible advice, and in extreme cases encourage self-harm, but these artificial companions are now insinuating themselves into our work and families and leisure, even offering themselves as replacements for our social or romantic partners,” notes Cass. “Teachers and students alike lament that the tools are undermining the foundations of the educational process; research is pointing strongly toward catastrophic declines in academic performance.”

There’s also the projection that AI will leave us all jobless — unless it kills us first, as tech insiders keep warning us.

In the meantime, they’re trying to force down the throat of every community at least one energy-guzzling, perpetually buzzing, water-hoarding data center.

Refocusing AI?

Cass makes the astute observation of how AI could improve our lives, if only the technology would be focused in that direction. “To really change health care for the better, A.I. must also make it possible for doctors and nurses to spend more time treating patients instead of dealing with administrative work,” says Cass. “It must manage health records more efficiently and eliminate the infuriating hassle of insurance paperwork. And it must help people run and interpret at-home diagnostic tests.”

AI companies can also show small businesses how their technology can help efficiency on a factory floor. They can coach builders, mechanics, and repairmen on how to use AI to help diagnose and solve problems. “The goal should be to make jobs better and more productive, not to make them disappear,” Cass suggests. And when jobs are eliminated, “A.I. companies serious about helping workers should establish large endowments to help people who need support during a transition and provide funding for the employers who then hire and train them.”

Justified Opposition

But these are not the experiences people are noticing. No doubt, AI is indeed providing some benefits. It’s answering phone calls for small businesses, it’s providing accounting services, it’s answering emails, it expedites research and creates digital designs. It’s even spotted emerging diseases far in advance. But what stands out to people are the negative aspects, which far outweigh the benefits. This technology is used to spy on us more efficiently, to manipulate us better, to help carry out strikes on schools in countries we never wanted to be at war with in the first place, and, now, to track us and perhaps even accuse us of crimes we’ve never committed.

The irony is that the predictive surveillance steps that Anthropic and other AI companies are building to ward off threats will likely only invite more criticism, more threats, and more attacks.

So, where is this ultimately heading? If people are already firebombing Sam Altman’s house, already destroying Flock cameras, already stalking AI employees, and, apparently already trying to kill AI execs — what will people do when they learn that they’ve already been deemed guilty before ever committing a crime?