Mayorkas: Government, tech are on "right path" to address AI risks
Add Axios as your preferred source to
see more of our stories on Google.
DHS Secretary Alejandro Mayorkas at Axios' AI+ Summit in Washington, D.C. on Nov. 28. Photo: Eric Lee/Axios
The Department of Homeland Security's work on AI has gone into "hyper-drive" as hackers start leaning more on generative AI in their attacks, Secretary Alejandro Mayorkas said Tuesday at Axios AI+ Summit in D.C.
Driving the news: During an interview with Axios' Ina Fried, Mayorkas said DHS is juggling several AI priorities — including implementing AI tools in its own criminal investigations and defending against the cyber threats AI is bound to accelerate.
The big picture: DHS is approaching its AI priorities in the same way it has approached other cybersecurity and national security concerns — through public-private partnerships.
What they're saying: "We are on a right path of a private-public partnership," Mayorkas said.
- "In my discussions with companies that are in the business of the development of AI and its advancement, I have certainly seen quite immediately a social consciousness — an understanding of its potential to do great, an understanding of the risks involved and a very significant desire to advance one and to manage the other," he added.
Details: Mayorkas also said that the department is already using AI tools to help with investigations involving fentanyl and missing children.
- With fentanyl, U.S. Customs and Border Protection is using artificial intelligence to detect unusual traffic patterns among vehicles crossing the U.S.-Mexico border and using that data to quickly identify those who may be illegally bringing fentanyl and other narcotics into the country.
- In missing children cases, Mayorkas said that DHS investigators have used AI to generate images of how a missing child would look years after they disappeared to better identify them.
Between the lines: DHS is at the forefront of implementing the long implementation to-do list coming from President Joe Biden's artificial intelligence executive order.
Yes, but: Much of the success of DHS's work depends on how well other stakeholders educate the public about the ways AI will be used — and the misinformation it can create.
- "The politicization of some of our core security responsibilities is extraordinarily worrisome," Mayorkas said. "When I speak of digital literacy, I speak of some very fundamental education programs and that requires a collective action."
Go deeper: Former Google CEO: Companies' AI guardrails "aren't enough" to prevent harm
