Homeland Security’s Pre-Crime Push

Snitch fund and AI vie to stop attacks like Boulder

Boulder attack suspect being arrested

The Trump administration is requesting Congress double a homeland security fund to pay off confidential informants, according to the detailed budget documents released Friday. It is the first time in a decade that the fund has been increased. 

The move is part of a broader push pitting traditional against cutting edge methods to prevent attacks before they happen, a goal enjoying newfound urgency following the Molotov cocktail attack on pro-Israel marchers in Boulder, Colorado yesterday.

“It is a whole new host of threat vectors you have to go after,” said Paul Mauro, the former chief of the NYPD’s intelligence arm, responding to the Boulder attack. Not very long ago, Mauro told Fox News, the nation’s domestic intelligence fighters were worried about “Al Qaeda having a sleeper cell here and ISIS weaponizing people over the internet.”

Now, Mauro continued, “the incubators are on campuses and in high schools and online” — in other words, everywhere. 

Preventing an attack before it occurs is a paradigm born out of 9/11. The idea was that a combination of police and intelligence work can be applied to stop plotters from committing crimes in the first place. This is the holy grail of domestic homeland security. But it hasn’t been particularly successful, with politically-motivated assassinations and related acts of violence on the rise.

As the Department of Homeland Security’s Fiscal Year 2026 budget request reveals though, the traditional methods of infiltrating groups, while on the rise, today constitutes a small part of the terror hunter’s arsenal. Artificial intelligence is garnering far more investment with the goal of building a surveillance system that will detect individuals planning such attacks, either in their social media affiliations and postings, or in the patterns of their communications.

Further research and development will allow the federal government to gather and process more intelligence than ever before.

Homeland Security’s new budget request hints at efforts being undertaken on a larger scale, alluding to “a unified National effort to prevent and deter terrorist attacks” led by the Department’s Chief Artificial Intelligence Officer.

The Homeland Security agency requesting a doubling of its fund to pay informants is Immigration and Customs Enforcement (ICE), best known for its deportation work, which of late has included raids into businesses and even restaurants to root out people residing in the U.S. illegally. The new informant funding will be used to infiltrate those tied to human smuggling networks as well as gangs and cartels. 

ICE’s FY26 budget request

ICE contains the largest federal law enforcement agency most people have never heard of: Homeland Security Investigations (HSI). The 7,000-agent strong organization has a broad mission to investigate everything from financial crimes to “national security threats.” 

While ICE’s informant budget request is the first increase since at least the Obama administration, HSI has used informants for years. An HSI training guide notes the role informants play in its “most significant investigations,” adding: “Confidential informants can be the eyes and ears in places that would otherwise be inaccessible to an HSI Special Agent.”

Informants are often used to penetrate target organizations in order to feed intelligence back to law enforcement about their activities. The government doesn’t say how many snitches it has. Customs and Border Protection, another agency of homeland security, alone has over 150 special agents handling its  informants, according to a recent government report, implying its informants must be in the thousands. But using informants is labor intensive, slow, and not always reliable.

Screenshot of HSI training guide

It is also a method of police work that isn’t always up to the challenges of dealing with “lone wolves” or non-organized crimes. 

That’s where AI comes in.

For years, social media “monitoring” on the part of the government has been on the rise, particularly in the war against ISIS. From foreign fighters moving around the globe to ISIS recruiters and propagandists communicating, social media has proven a rich source for intelligence people seeking law enforcement and targeting leads. The flood of communications, though, overwhelmed the capacity of analysts, just at a time when artificial intelligence was showing promise as a way of triaging huge amounts of data.

Now, federal government investments in AI are going into overdrive. Projects, open and secret, are seeking ways of untangling the world of social media, this time with an eye to ingesting anything and everything, looking for a haystack and then a needle, all in real time.

“We don’t want the smoking gun to be a mushroom cloud,” National Security Advisor Condoleezza Rice famously said after 9/11, suggesting that the next attack by al Qaeda would come from a weapon of mass destruction.

In hindsight, it wasn’t a threat based upon probability; it was derived from Rice’s own background in Cold War nuclear weapons and the arms race paradigm, a paradigm that Cheney and Rumsfeld and others also slotted into because that’s all they knew.

Beyond WMD, investigators approached al Qaeda as a crime family, even using such language, because organized crime was their experience, and the methods that were applied to eliminating an organized crime family sufficed for two decades.

Then the focus turned to people “inspired” by foreign terror groups in the ISIS era, and all of the programs to attack the radicalization process and “counter extremism.”

Now, under Trump, but not only Trump, it’s the ideology of anti-semitism and hate crimes, one that expands the battleground to what everyone thinks and says. 

This week, former New York City Comptroller Scott Stringer announced a new partnership with the Anti-Defamation League to come up with AI-methods to “predict and prevent” hate crimes in America, the battlecry that fits with the anti-hate industry’s worldview. 

“My new initiative would use advanced monitoring tools to scour social media posts following during or after international emergencies,” Stringer said.

Yes, ICE is doubling its walking-around money to recruit informants among its long list of enemies; but AI? That’s coming after all of us.

— Edited by William M. Arkin