Alejandro Mayorkas held what’s been called “the hardest job in Washington” during a particularly turbulent time in U.S. politics. As Biden’s Secretary of the Department of Homeland Security, he endured an impeachment by House Republicans, a fentanyl crisis, a string of natural disasters, and a number of terror attacks on home soil. But he also ran the sprawling department during a major technology shift with generative AI, and was charged with leveraging the technology to make DHS more efficient and responsive. Politically, Mayorkas may be remembered as an immigration lightning rod (Mayorkas himself immigrated to the U.S. from Cuba as a child), but within the DHS he’ll also be known as the guy who mixed AI with bureaucracy.
In an exit interview with Fast Company during one of his final days at DHS, Mayorkas gives his highlight reel of the department’s AI work, which included working with tech sector partners including Sam Altman and Jensen Huang, and recruiting AI engineers in Silicon Valley. The interview has been edited for length and clarity.
The generative AI boom came during your term as secretary. How well did DHS respond to this big technology shift?
I will say quite candidly we responded superbly, and I say that giving credit to extraordinary people here in the department. We moved swiftly and expansively, and we have led the federal government in embracing the potential of AI in all regards: to harness its potential to advance our mission, to understand the potential for its malevolent use and begin to take steps to protect the homeland and most particularly critical infrastructure against that challenge, to build a workforce that is capable of both harnessing and securing AI, building partnerships with the private sector, and promulgating policies to govern the use of AI into the future.
Having a primary role in different regards that are codified in presidential executive orders, we have done so very much. We moved very quickly in developing our AI road map early in 2024. We issued policies and procedures governing our use, we assembled the AI Safety and Security Board that I chair with luminaries from throughout the AI ecosystem, and we created the position of chief AI officer [currently held by Eric Hysen]. We hired 49 experienced technologists to form our AI Corps. I could go on and on.
Are there any moments that you remember being particularly joyful, like seeing this technology applied and making a difference on the ground?
I can identify a number of moments, but in terms of seeing AI in action, it is remarkable to see us having trained a machine to simulate a refugee applicant in order to provide training for our refugee officers. The training yielded not only a machine that could respond to the officer’s questions substantively in terms of country conditions and what the machine applicant lived through, but also trained the machine in how a refugee applicant would answer the questions given the trauma endured. For example, training the machine to be reticent in providing certain details—it’s extraordinary. I cite that because I’m a political refugee myself. I came to this country as a one year old, but certainly my parents shared with me their refugee experiences and so it was quite resonant.
I understand that there are now 158 AI applications in use across the department, including 29 systems that are directly impacting immigration processing and document processing.
We are looking at our work across the spectrum throughout the department and assessing where we could deploy AI to advance that work. There’s no stone left unturned in the exploration of the potential for AI in our work.
It seems like there’s a lot of friction in the processing of people coming to this country. It requires gathering a lot of information and documentation. Is that something AI might solve?
It should be used to drive transformative efficiency. I will tell you when I entered the Immigration Agency, U.S. Citizenship and Immigration Services in 2009, it was purely paper based. I mean, there was a small amount of online capability, but generally aerospace. And this is the 21st century and so we have modernized our processes considerably. Now we have AI which could take that to a very different level, and when you speak of friction, I am thinking of bureaucratic friction, if you will, and the elimination of that. The political friction—that would be magical if AI could address that.
Your chief AI officer, Eric Hysen, and his team produced a Generative AI Public Sector Playbook that contains learnings from everything DHS has built with AI. How do you hope the document might be used in the future?
We have published a number of seminal documents because our focus has been on institutionalizing this commitment to the potential of technology writ large, and obviously AI can be the most transformative technology in a long time. We thought it both prudent and important to publish documents that will help institutionalize this commitment and hopefully secure the longevity of this commitment.
This year you did a lot of work to bring together the Artificial Intelligence Safety and Security Board, which includes an impressive lineup of names that we’re familiar with in AI circles. Is that board going to survive through the next administration?
That is a question that should be directed to the incoming administration. It is certainly my hope that it continues. We have, in collaboration with that board, produced a groundbreaking framework for the safe and secure deployment of AI in critical infrastructure, and that board is apolitical. As you noted, it is constituted by luminaries in the tech industry, in critical infrastructure, in civil society, and in public office.
And the board produced the Framework for the Safe and Secure Deployment of AI in Critical Infrastructure, released in November.
It was a significant investment of thought and time and energy on the part of this department, me, and a number of people here, and the board members and their respective teams.
Turning focus onto the recruiting you did to staff the AI Corps, I was especially interested to hear that you personally went to the West Coast to interview candidates. Through that experience, what did you learn about how Silicon Valley people think about government service?
What I did in my recruiting efforts in Silicon Valley and elsewhere in the country was share with them what we do, the impact on people’s lives in a myriad of ways, and invited them to be a part of it. To not think of government from a 50,000-foot perch of rhetoric, but rather I explained to them what we do that impacts people’s lives, makes people’s lives better, and what an incredible privilege it would be for them, and how fulfilling and inspiring it would be for them to be a part of it. We were recipients of a wave of applications, thousands if I am correct.
That applies to the way you talk about AI to DHS people internally too, I imagine.
Yes, and I must say, our workforce is very excited about the potential of AI. We want to be clear that we are leaning forward significantly with appropriate guardrails to ensure that our use of AI is responsible, that it is respectful of our commitment to civil rights and civil liberties and privacy interests.
I know responsible AI has been a key concern but, you did receive a letter from a number of immigration civil rights groups in September saying that DHS is fast tracking these AI applications without enough regard for transparency and civil rights. Did you meet with these folks and what was the result?
Yes, we met with them and sent a response. The letter has mistakes of fact and also contains a misunderstanding of what we do and don’t do, and so the letter we set forth in response, I think will lay it out quite clearly. There are a lot of mistakes and misunderstandings in that letter.
They were particularly concerned about ICE using biometrics data with AI. Can you comment on that?
I’ll refer to the blog post and the responsive information because we dealt with each.
I know you’ve been speaking with both [DHS secretary nominee] Governor Kristi Noem and the Trump transition team. Was AI a part of those discussions?
It’s not for me to speak of the conversations I’ve had with Governor Noem or that our team has had with the transition team. I will say that I have spoken with Governor Noem a number of times and have been both impressed and appreciative of those conversations.
Some of the DHS AI initiatives grew from the White House’s executive order on AI. Some Republicans in Congress have already expressed an interest in reversing the EO. Could that derail some of the department’s efforts or endanger DHS funding for AI work?
I don’t know what actions will and will not be vis-à-vis the executive orders. But I will tell you that I am optimistic that the incoming administration will continue to employ technology to advance government missions. I am quite optimistic about that.
Autentifică-te pentru a adăuga comentarii
Alte posturi din acest grup
Opinions can be shaped very quickly, and grudges last a long time. Donald Trump knows that better than most. Joe Biden ought to, but seemingly doesn’t.
How else to explain the naivety of
A SpaceX
In a unanimous ruling, the Supreme Court voted Friday to uphold a law that will require
President Joe Biden won’t enforce a ban on the
When a truck breaks down, its driver knows what to do. They pull over to the side of the road, put on their hazard warning lights, and lay out safety triangles to warn other drivers about the risk
In the great AI spending spree, does ROI matter?
That’s a question that every tech-related company should be asking themselves. The right answer could save businesses a lot of grief whil
As we look ahead to 2025, the financial services landscape will be shaped not by AI as a standalone innovation, but by how it enables smarter, faster, and more personalized experiences. Here are t
