![](\"https:\/\/www.newsbeep.com\/us\/wp-content\/uploads\/2025\/10\/UMDAI02Resized-480x320.jpg\"\/)
\n Students who won the University of Maryland’s XFoundry Xperience competition created an AI program that can detect a weapon inside a school building within about three seconds of when someone pulls it out.
\n(WTOP\/John Domen)<\/p>\n
Last month a group of U.Md. students who formed a company called DefenX that uses AI to detect weapons in schools were announced as the first winners of the XFoundry Xperience competition.<\/p>\n
\n Students who won the University of Maryland’s XFoundry Xperience competition created an AI program that can detect a weapon inside a school building within about three seconds of when someone pulls it out.
\n(WTOP\/John Domen)<\/p>\n
![](\"https:\/\/www.newsbeep.com\/us\/wp-content\/uploads\/2025\/10\/UMDAI02Resized-780x520.jpg\"\/)
\n Students who won the University of Maryland’s XFoundry Xperience competition created an AI program that can detect a weapon inside a school building within about three seconds of when someone pulls it out.
\n(WTOP\/John Domen)<\/p>\n
For students born after the Columbine High School shooting in 1999, active shooter drills have become a routine part of growing up. Now, as young adults, many are questioning why this has to be their reality at all.<\/p>\n
Solving that problem became the first goal of the University of Maryland\u2019s xFoundry Xperience<\/a> competition, which brings together students from different majors and disciplines \u2014 STEM and non-STEM \u2014 to solve complex, societal issues.<\/p>\n The program launched in early 2024<\/a>, and last month, a group of students who formed a company called DefenX<\/a> were announced as the first winners. They\u2019ll be getting a $250,000 prize to help launch their company.<\/p>\n Using camera systems already in place at schools, the students honed and trained their AI models to detect a weapon within about three seconds of when someone pulls it out. The accuracy rate is over 90% \u2014 better, they claim, than some of the products already on the market.<\/p>\n \u201cAs soon as our system detects that a firearm is in place, we\u2019ll be able to \u2026 also pinpoint the shooter\u2019s exact location within the school, and send alerts seamlessly,\u201d said Smithi Mahendran, a computer science major and one of the students involved with DefenX.<\/p>\n \u201cWhat we would do is track the shooter, detect them, and communicate with all relevant authorities such as schoolteachers, students and law enforcement in real time,\u201d she added. \u201cAnd then lastly, we\u2019ll also provide you with end-to-end support, which includes mental health resources, a chat bot that allows you to review incident footage data, among many other things.\u201d<\/p>\n Over the course of about 18 months, Mahendran said the AI models were able to detect weapons in all of the situations that can be encountered in school, from crowded hallways to dingy lighting. And it can differentiate between a firearm and an umbrella within that crowd.<\/p>\n \u201cAccording to our test, it\u2019s about three seconds \u2026 totally for detection,\u201d she said.<\/p>\n The xFoundry program was launched by Amir Ansari, who dubs himself a serial entrepreneur with dozens of patents to his name. He said growing up with active shooter threats and drills has had a lasting impact on the students in the program.<\/p>\n \u201cThey see themselves now as part of the solution set,\u201d Ansari said. \u201cAnd they saw themselves in a position to do something about it.\u201d<\/p>\n An on-screen map marks the shooter\u2019s gun<\/p>\n DefenX students tested the system inside a building on the university\u2019s campus.<\/p>\n In a demonstration video, when the cameras detect a gun, a clear photo is generated within the program and a map of the school\u2019s hallways lights up with a red dot marking the person with the firearm. It can track them through the halls and into rooms. They hope to install the program inside some area schools to conduct a pilot study, and have the system on the market next year.<\/p>\n \u201cOnce it detects the shooter, it\u2019ll be able to track them and see how they move and then update the coordinates accordingly,\u201d sai Srinidhi Gubba, a computer science major.<\/p>\n The red dot is replaced by a yellow one that marks the last spot the person with a gun was seen if they ever move out of a camera system\u2019s vision.<\/p>\n \u201cIt\u2019s really disheartening for us to just even think about how we have to prepare for situations like this,\u201d Gubba said. \u201cBut with our solution, we think we can make it, at least a little bit, more safe for schools.\u201d<\/p>\n While Gubba and Mahendran were involved in the technical side of things, other students in the competition were focused on other areas, including the ethical side of the AI program. For instance, no facial recognition data is kept by the AI system.<\/p>\n In addition to all that, DefenX checked the box for the part of the competition that made sure the products developed would also be affordable for school systems to actually budget for.<\/p>\n \u201cThese are very difficult requirements to make happen,\u201d Ansari said. \u201cBut the students figured out a way to do this from a technology perspective, from a user interface perspective, from the operational requirements that they have to have so the business can still thrive while they\u2019re providing these amazingly low-cost solutions for every school in the country.\u201d<\/p>\n The next cohort of students are already working on the second xFoundry competition, which focuses on students\u2019 mental health concerns. And while they\u2019re being given a lot of freedom to come up with a solution that will help, the goal is to create something that students will use on a continuous basis to address their mental health.<\/p>\n \u201cWe have some of the smartest people that have ever come through colleges with all the resources that these students have,\u201d Ansari said. \u201cIt\u2019s a good idea, in my opinion, to put them in the pathway of these problems and allow them the freedom to find a solution that works. More often than not, I believe they\u2019re going to find real, clever, awesome solutions that we can all benefit from.\u201d<\/p>\n