Reining in Rogue AI: Stevens Students Lasso First Prize and $10K in Homeland Security Design Challenge

A student team from Stevens Institute of Technology took home top honors and a $10,000 prize at the Department of Homeland Security (DHS) Designing Actionable Solutions for a Secure Homeland (DASSH) student design challenge. The winning Stevens trio, dubbed the DASSHing Ducks, conceived the LiDAR LASSO, an innovative security system designed to stop autonomous vehicles in their tracks while letting human drivers pass.

Hosted remotely at Arizona State University, the 48-hour DASSH 25: Security in an AI World brought together 160 students from 15 universities to tackle real-world challenges involving artificial intelligence and national security, with the goal of turning student ideas into viable defense tools. Teams chose one of three problem statements, each focused on protecting the nation from emerging AI-enabled threats. 

Unlike traditional hackathons, this competition pushes students to think beyond coding. They are encouraged to blend AI, engineering and creativity while collaborating with peers, academic mentors and industry experts on innovative solutions to pressing homeland security concerns.

"This is not just another academic challenge," noted Julie Brewer, acting undersecretary for the DHS Science and Technology Directorate, at the DASSH challenge. "It is an opportunity for all of you to design, build and implement innovative solutions that future generations of homeland security practitioners can use. The ideas, concepts and solutions you will put forward this weekend matter."

Creating fake barriers to thwart real threats

The interdisciplinary DASSHing Ducks team chose to respond to the problem statement: How do we mitigate the potential for AI-enabled attacks on critical infrastructure?

With recent vehicle-based ramming tragedies in mind, they developed a tool designed to detect and disable potentially dangerous autonomous vehicles using LiDAR (light detection and ranging) spoofing. 

Current countermeasures, such as physical barriers and signage, have limitations: they’re costly, they’re not portable and they block all traffic, including legitimate vehicles. The Ducks floated a lot of ideas, finally narrowing in on their winner. 

Like a digital cowboy’s rope, their LiDAR LASSO throws out laser-based illusions to stop unmanned vehicles before they can do harm. Human drivers can’t see the laser obstructions, which are outside the visible light spectrum, so they just keep going. It can also alert law enforcement when a suspicious vehicle stops for the false obstacle.

"It’s like creating an invisible force field that only driverless vehicles can see," explained team member Andrew Vallent, class of 2026, who is majoring in Naval Engineering in the Department of Civil, Environmental and Ocean Engineering.

The system is modular and portable, with easy mounting on existing infrastructure, so it’s ideal for protecting high-risk sites without disrupting traffic. The estimated cost of a unit is just $585, and it’s scalable for use to secure stadiums, power plants, airports and other high-security areas.

Don’t sleep on this smart solution

The team worked around the clock during the 48-hour sprint, squeezing every ounce of creativity and energy into their design. 

"We didn’t get much sleep either night," admitted Gabriel Pust, Class of 2027, who is pursuing his combined bachelor’s and master’s degree in Naval Engineering and Ocean Engineering. "But the thrill of the experience – coupled with short deadlines – sparked a sense of indefatigability that carried us through."

The team found the most difficult part to be determining how AI would not simply enhance, but truly enable an attack. Their diverse backgrounds in engineering, finance and security – along with a keen sense of humor – offered a unique edge.

"It felt like we were treading the line between innovative and ridiculous," Pust said. "Trying to come up with a way to keep our idea practical to implement while still completely novel was entertaining."

The experience also helped the students grow as problem-solvers and keen observers. 

"We learned how to create and pitch a product for a client, and design with purpose," said Cian Gahan, Class of 2027, who is studying Quantitative Finance. "Plus, it made us realize AI has already infiltrated more of daily life than we thought."

Mentor and Stevens alum Christopher "Kip" Francis ’11 (B.S.) ’13 (M.S.), a research engineer at the Davidson Laboratory and a veteran of DHS events, praised their creative energy and instincts.

"I expected I’d spend half the weekend explaining how DHS thinks, what the lingo is and what the major concerns might be," Francis said. "But these students surrounded the problem and closed in on a solution, all while displaying knowledge about the homeland security world I didn't know they had. I mostly sat back, encouraged them to trust their instincts and enjoyed the show as they worked toward their creative solution. It reinforced to me that Stevens continues to attract the best and brightest."