One of the most heated debates of our time — border security — is getting an AI-powered twist.
Congress is looking to fund machine learning, biometric scans, and drones to keep “potentially dangerous people” out. Hundreds of millions of dollars are appropriated to technologies in partnership with private companies. And the projects raise questions about effectiveness, inherent biases and the potential to erode civil liberties.
Border tech is nothing new.
U.S. border policy has been on a gadget binge since the 1990s. The Biden administration leaned into “smart border” initiatives over concrete barriers, channeling over a billion dollars into AI-powered analytics, advanced drones, and next-gen surveillance towers.
If you live in Southern Arizona, you might’ve seen the surveillance blimps hovering overhead.
The Biden administration took a somewhat cautious approach to implementing AI technologies, developing frameworks and roadmaps emphasizing “responsible” use while “protecting individuals’ privacy, civil rights, and civil liberties.”
Donald Trump's return to office brought a sharp pivot in AI policy. On January 23, 2025, he signed an Executive Order rolling back Biden-era AI regulations, calling them "harmful" barriers to innovation. It’s unclear how this will shape AI’s role in border technology, but with Elon Musk now playing an active part in the administration, the emphasis on deregulation and rapid innovation is already evident.
Now, a bipartisan bill in Congress is making another push at securing the border with high-end technologies.
H.R.7832 - Emerging Innovative Border Technologies Act, would have required the U.S. Customs and Border Protection (CBP) and Department of Homeland Security to develop a plan for using advanced and disruptive technologies at all international borders and ports.
The House gave the bill its nod last September, but the Senate let it quietly fade from the legislative spotlight. It resurfaced this month with fresh momentum in a more tech-friendly Congress.
The bill doesn’t mandate specific systems, but it does outline categories of technologies that CBP should consider exploring, piloting, and potentially integrating into official programs.
Some of the technologies include:
Unmanned aerial systems and related technology
Tower-based surveillance technology
Tunnel detection technology
Supporters like the bill’s Democratic co-sponsor, U.S. Rep. Lou Correa of California, say that digitizing border security with machine learning will help keep communities safer by detecting drugs and stopping trafficking before it crosses onto American soil.
“Border security means keeping drugs and other negative elements away from our communities—and cutting-edge technology that is already available for commercial use gives our hard-working officers the tools they need to keep us safe,” Correa said.
Meanwhile, civil liberties groups argue the bill’s broad scope could mean more intrusive surveillance, algorithmic bias in facial recognition systems, and a ramp-up of warrantless data collection — especially in already over-policed border communities.
Tucson’s Democratic U.S. Rep. Raúl Grijalva has emerged as a leading skeptic of DHS’s use of AI in immigration. He spearheaded the December 2024 letter urging DHS to halt AI-driven immigration decisions until proper safeguards are in place.
“There’s a really delicate balance between pushing for innovation and efficiency and automating life-or-death decisions, especially when it comes to immigration enforcement and adjudication,” he said in a letter to the Biden administration in December. “Further integrating AI into our immigration system without the proper guardrails poses a great risk to people’s civil liberties.”
As a longtime advocate for immigrant rights and border communities, Grijalva is concerned about due process for migrants and whether DHS’s AI complies with federal standards and he continues to press for transparency from DHS.
Arizona has been at the forefront of AI-driven border security research, and if the bill does pass, Arizona universities could see a windfall of AI-driven border security research funds.
Arizona State University’s Center for Accelerating Operational Efficiency (CAOE), for example, has worked on AI-assisted border screening to optimize wait times, improve security, and study human-AI collaboration.
Meanwhile, the University of Arizona’s BORDERS Center developed AVATAR, an AI-powered deception detection system tested at border crossings and international airports.
But the overall success of these initiatives remains unclear — Hank wrote about that AVATAR system 14 years ago, and it still isn’t in widespread use today.
Over the past few years, federal spending on border technology has skyrocketed. Hundreds of specialized companies have entered the space, competing for lucrative contracts.
And with big money and new technology comes government waste and corporate grifters.
One major example of waste in AI-driven border tech is CBP One, a mobile application designed to facilitate asylum applications and border entry processing. Initially launched in 2020, it was funded and expanded under the Biden administration as a key tool in streamlining legal border crossings. The app had major flaws — technical issues, bias in facial recognition, frequent crashes — and Trump shut it down on his first day in office.
With no plan to replace CBP One, its shutdown exemplifies how political swings can turn multi-million-dollar investments into digital dust.
The Government Accountability Office (GAO) has long warned about the dangers of investing in AI tools without oversight. And as AI becomes further embedded in border enforcement, concerns over data privacy, civil liberties, and for-profit surveillance grow.
As bills like H.R.7832 make a comeback, it is important for policymakers to take a deeper look at the evaluation and long-term improvement plans of systems that are being adopted. As we’ve seen happen in several places, it is too easy to spend billions of dollars and have nothing to show for it.
No opportunity without risk: Vice President JD Vance warned global leaders against over-regulating and slowing down the growth of AI technologies. In his speech at the third annual “Artificial Intelligence Action Summit” in Paris, the VP emphasized that the focus of AI policy should be on seizing its opportunities rather than regulating it, and he pledged that the U.S. would continue to be the global gold standard in AI by ensuring innovation, fostering international partnerships, and protecting the AI ecosystem from overregulation. He also said American AI will be free from ideological bias and will not be used as a tool for censorship or suppression of free speech.
“I really believe AI will facilitate and make people more productive. It is not going to replace human beings,” Vance said. “I think too many of the leaders in the AI industry when they talk about this fear of replacing workers really miss the point: AI, we believe, is going to make us more productive, more prosperous, and more free.”
They’re taking our jobs: Fortune magazine tells the story of Valerie Gills, a former Phoenix hotel receptionist whose job was replaced by AI chatbots, noting that Latinos are overrepresented in jobs at high risk for automation, per the ACLU Policy Institute. After getting laid off, Gillis got an internship as a blog editor, and is now looking for tech programs to upskill into the workforce of the future. Meanwhile, ASU built an AI-powered sandwich-making robot that can make “five different types of sandwiches.” It’s one of the many innovations coming out of ASU’s School of Computing and Augmented Intelligence, ABC15’s Adam Klepp writes.
So much innovation: Elsewhere at ASU, researchers are using AI to predict when your chances of catching Valley Fever are highest, KJZZ’s Hailey Jimenez explains. Also, ASU and the Arizona Department of Emergency and Military Affairs are using a research grant from the Arizona Board of Regents to build an AI-powered emergency detection system, which they hope will be able to predict and detect wildfires, floods and other threats to infrastructure, KTAR’s David Veenstra reports.
Not all of it works: An AI-powered gun detection system that a school district in Tennessee bought didn’t work, per States Newsroom’s Page Gross. The school district paid $1 million for the Omnilert gun detection system, but the cameras on campus weren’t positioned in a way that the system picked up the shooter’s gun before he opened fire, killing one classmate and wounding another. The system did detect the guns of police who responded.
Raw power: Lawmakers in Utah want to build new nuclear power plants to triple the state’s energy output by 2050, hoping to become a power leader in the power-hungry age of AI, Pluribus News’ Reid Wilson reports. Senate President Stuart Adams introduced a package of bills to encourage building nuclear, mining more natural gas and expanding a geothermal plant, saying it’s a matter of national security. Meanwhile, the Arizona Daily Wildcat’s Sam Parker and Kanishka Chinnaraj break down exactly how much water and power it takes for ChatGPT to crank out those late-night college essays.
“The country that’s going to control AI is going to control the world, both economically with goods and services and militarily with weapons systems,” Adams said.
Money for free: If you’re using AI to write your grant applications, GovTech has some tips and tools for you to check out.
Perplexity AI is an AI-driven search engine that’s kind of like what would happen if Google and ChatGPT had a baby.
It provides quick and reliable answers to your ill-formed search queries, backed up with the links and citations you need to verify the answers or dig further into the rabbit hole.
It also features a pretty decent "Discover" section, where AI curates and summarizes the latest advancements across various fields.
