Every two seconds, someone in the United States needs a blood transfusion. Their reasons for needing blood vary—surgeries, cancer treatment, chronic illnesses, and traumatic injuries are typical—but all of them are dependent on the millions of Americans who donate blood each year.

It’s a dizzying logistical challenge to coordinate the collection, delivery, and storage of donated blood to ensure that it’s available to patients when they need it, and since 1948, the American Red Cross has been largely responsible for keeping that blood flowing. Today, the Red Cross supplies 40% of all donated blood in the United States and has a near-perfect record of ensuring that patients receive the blood they need. Most blood shortages have been local and short-lived—until last year, when the Red Cross announced the first national blood crisis in the organization’s history.

The crisis was driven by the lingering effects of the COVID-19 pandemic, which led to a 10% drop in the number of blood donors, as well as operational challenges such as canceled blood drives and staffing shortages. Although the crisis has improved somewhat over the past year, local shortages continue to roll through U.S. cities, and in July the Red Cross announced it had received 50,000 fewer blood donations than it needed in the preceding two months.

The Red Cross has a massive amount of donor data that it can use to select optimal donation sites, based on factors such as the density of past donor populations, the likelihood that a donor will return, and the optimal times for collection. It’s too much data for a human to comprehend, but artificial intelligence allows the Red Cross to extract actionable recommendations from the dataset. This type of AI system is technically complex, and the stakes are high. But with the help of the students, faculty, and industry partners who make up Penn State’s Nittany AI Alliance community, the Red Cross hopes to make its vision for AI-driven donor site selection a reality.

This fall, three teams of Penn State students participating in the Nittany AI Alliance Advance program—a paid internship in which students build AI tools for social impact—began working with the Red Cross to build AI tools designed to help it increase blood donations. Each team will focus on a different area of the problem: finding ideal locations for blood drives; predicting the number of young donors who will show up at a site; and predicting the behavior of donors, such as whether they will show up to their donation appointment on time. The goal of each team is to produce a working, AI-driven prototype that the Red Cross can build upon to solve its donation challenges. Throughout the project, they’ll be supported by Penn State faculty and two Nittany AI industry partners—Lockheed Martin and Accenture—that are financially supporting the initiative and providing expert insight on project management and techniques for synthesizing data from multiple sources.



animated conceptual illustration of an arm receiving a blood transfusion and color squares by Stuart Bradford


The Red Cross has its own team of expert data scientists capable of building these types of tools internally, but the organization must balance a wide range of competing demands on its budget and staff. It’s difficult to justify allocating scarce resources to a project unless it shows signs that it will succeed. This is where Nittany AI comes in.

“Exploratory analysis is a big deal because you want to know that the solution using this data is going to solve the problem,” says Michael Bryan ’19 MAS Sci, a Red Cross AI consultant and the organization’s liaison with Nittany AI. “By working with Penn State, we’re opening ourselves up to finding questions, solutions, observations, and analytical results we may not have otherwise thought of or found. It really benefits the students at Penn State, and it benefits the Red Cross. It’s a wonderful win.”

This vision—the kind of close collaboration among students, faculty, industry experts, and nonprofits to use AI for social good—is what Nittany AI has been building toward for years. At a time when fears about the unethical use of artificial intelligence are at an all-time high, Nittany AI’s work showcases how these new systems can lead to positive outcomes. The Alliance has also built a model for innovation in higher education that helps students acquire a variety of future-proof skills through hands-on, real-world experience.


For most of his career, Daren Coudriet has worked a bit further in the future than the rest of us. After finishing his undergraduate and graduate degrees, Coudriet ’87 Eng, ’92 MBA Bus, ’92 MEng Eng flung himself into the technology sector during the ’90s dot-com boom: He advised utilities on how to use the web to cope with deregulation and helped build the first websites for media giants such as The New York Times, The Washington Post, and Standard & Poor’s. What united his various ventures was his knack for helping organizations proactively adapt to rapid technological change.

“I built a career [on] helping businesses modify their processes so they can ride the wave of disruption instead of getting crushed by it,” says Coudriet, who since 2016 has served as executive director of innovation and the Nittany AI Alliance at Penn State Outreach. “I was the guy who understands both business and the affordances of new technologies.”

In 2015, Craig Weidemann, then Penn State’s vice president for outreach and vice provost for online education, hired Coudriet as a consultant to explore how emerging technologies might affect the university. It was an open-ended assignment, but it wasn’t long before Coudriet narrowed in on artificial intelligence as the technology that would have the biggest impact. If Penn State wanted to prepare its students for the future, he knew the university urgently needed new ways for them to engage with AI.

“Look at the way the internet has changed the way we lived over the past 20 years,” says Coudriet, recounting his discussion with Weidemann. “AI is going to have 10 times the disruption in half the time. It could potentially disrupt the entire higher education model. I thought Penn State should just focus on AI and how that might be used to help it become more efficient and reduce costs.”

Weidemann agreed. Together, he and Coudriet hatched a plan for what would eventually become the Nittany AI Alliance. It would be one of 11 organizational units within Penn State Outreach, and it would be dedicated to two core goals: educating students about AI through hands-on projects and using those AI projects for social good. It was a mission uniquely suited to a university like Penn State.

“As the only land-grant institution in Pennsylvania, we have this heightened sense of responsibility to make sure that we’re doing all that we can through our resources to improve the public good,” says Larry Terry II, vice president for outreach. “We wanted to look at AI as a tool that we can use to benefit our communities.”

In 2017, Coudriet launched Nittany AI with the Nittany Watson Challenge. Sponsored by IBM, the challenge required entrants to build tools that could improve the student experience using the company’s Watson AI platform. Over eight months, student teams developed AI tools to help with common problems such as scheduling courses or finding jobs. Their projects were presented to a panel of judges, which awarded 10 teams $5,000 each in seed money to continue developing their prototypes. A final round of judging awarded five teams with $10,000 in prize money.

Among the winning entries were tools for evaluating transfer credits for prospective students and a virtual academic adviser. The competition validated Coudriet’s hunch that the best way to prepare students for the future was to have them build AI tools for real-world problems. But he also realized that a competition wasn’t enough to deliver on Nittany AI’s ambitions. He needed a way to engage students so that they could continue developing their skills and have the chance to complete projects that might last more than a few months. Coudriet also wanted a way to engage students who were outside of traditional AI-focused disciplines such as computer science and engineering; early on, he believed AI would affect everyone, so it was crucial to engage students studying liberal arts, business, and other “nontechnical” subjects.

“They may think they don’t have to worry about AI, but at a minimum they need to understand how it might impact their careers,” he says. “You [don’t] have to understand everything about the technology; you just need to know how to leverage it.”

To meet these objectives, Coudriet organized Nittany AI around three core programs: Inspire, Challenge, and Advance. Nittany AI Inspire would be an educational forum designed to engage students by discussing how AI is relevant to them. The annual Nittany AI Challenge competition would enable students to think outside the box and focus on creating working prototypes that solve real-world problems. And Nittany AI Advance would provide paid internships in which small teams of students develop an AI proof of concept for a nonprofit partner.

“One of our key principles is to meet students where they are,” Coudriet says. “The last thing they want to do is go to another event to hear someone talk, where they don’t see any connection to what they’re doing. Instead, we go to existing events and inject AI.”


When Christie Warren entered the second Nittany AI Challenge in 2018, the thought of winning never crossed her mind. Warren ’21 A&A was a sophomore graphic design major with no experience in AI or machine learning. When a friend asked her to join his team, she says, “I thought I might as well try it. What’s the worst that could happen?”

For its entry, Warren’s team built an AI-powered tool that generated a course list for students based on previously taken courses and the requirements for their major. As the team members knew from experience, planning courses could be a challenge for students, especially those with multiple majors or minors. Unlike many of her teammates, Warren didn’t have software development experience, so she used her design talents to create an easy-to-use interface for the app.

To Warren’s surprise, her team won—a victory that came with a $30,000 prize funded by university and industry support—and several of the judges, including senior university leaders and representatives from Google, IBM, and Microsoft, cited the interface as a decisive factor in their vote. For Warren, AI no longer seemed like an intimidating technology accessible only to technical experts. Along with her teammate from the previous year, Matthew Mancini ’19 Eng, Warren entered the Challenge again the following year, this time with a tool designed to automatically generate quizzes for students based on course material. That entry won, too.

“The success is entirely based on what you’re willing to challenge yourself to do and learn,” Warren says. “That was really inspiring and made me want to learn more outside of class.”

The Nittany AI Challenge also serves the crucial function of helping prepare and recruit students for its Advance program. Patrick Elisii ’23 Eng first entered the Challenge as a freshman, and although his team’s project—an AI for sorting recycling on campus—didn’t win, the skills he picked up by participating proved worthwhile. “I learned a lot my first time in how to run a team and how to create a project that has a real impact on what people are trying to do,” he says.

Elisii first worked with Nittany AI Advance as a sophomore on a project sponsored by Lockheed Martin to develop an AI system that could help find hikers lost in the wilderness. “We all learn how to code in our classes, but we typically don’t learn communication skills on top of the technical skills—one of the most important things companies look for when they’re hiring software engineers,” he says. “You have to learn how to speak to clients and how to ask the right questions. I learned how to take a problem and design a solution for it, rather than just taking an instruction set and completing it.”

The search for solutions to real problems drives every Nittany AI project. In 2020, the Southeastern Pennsylvania Transportation Authority (SEPTA) partnered with Nittany AI on a project designed to measure the impact of the COVID-19 pandemic on transit ridership. The project was short—lasting only one summer—but SEPTA was impressed by the students’ abilities, and in 2022, SEPTA pitched a more ambitious project aimed at solving a pernicious problem. Each year, dozens of people fall onto train tracks, for reasons ranging from inebriation to suicide; many of those falls could be prevented with proactive monitoring and alert systems. SEPTA hoped to develop a prototype AI system that could analyze security camera footage and automatically alert safety officers when it detected a potential fall.

“Once we knew this resource existed at Penn State, it became clear that it was something we could really take off with,” says Grant Engel, data policy manager at SEPTA’s innovation office. “We get to be creative and fulfill [a] need without stretching too thin, and we can impart knowledge to the students so they’re learning in a fun and cooperative way.”

Nittany AI Advance interns worked with SEPTA staff to develop a “fall detection” algorithm based on video data. In April, the team unveiled a prototype that could reliably identify when a rider fell; the system is not yet ready for real-world deployment, but Engel says SEPTA remains interested.

Prioritizing projects that impact Pennsylvania, Nittany AI has also partnered with Goodwill Keystone, which oversees Goodwill locations across the state. Each year, Goodwill stores receive millions of donated items, which must be sorted and priced. The sheer volume of items, which are processed manually, means many are incorrectly sorted and mispriced. A Nittany AI project launched in 2022 explored the use of AI to identify brands of shoes received by Goodwill and sort them based on the company’s inventory and sales patterns.

Andrew Gackenbach, chief retail operations officer for Goodwill Keystone, acknowledges the skepticism around AI but says that “working with Nittany AI has made it obvious that we need to wake up and pay attention to this, because it’s coming whether we want it or not. [With this project] we can create a very low-risk proof of concept and make a calculated business decision rather than a gamble.”


conceptual illustration of shoes, bar codes, and colored squares by Stuart Bradford




Corporate and nonprofit partnerships aside, one of Nittany AI’s priorities since its inception has been finding ways to leverage AI to help Penn State. In 2019, it launched its flagship Advance project with the Office of Admissions with the goal of improving the high school student application and evaluation process. Each year, the university receives around 150,000 transcripts from prospective students, but there is no standard transcript format, which greatly slows the analysis process while increasing both costs and the likelihood of mistakes. It was precisely the sort of problem that AI is well suited to solving.

Coudriet says the admissions office was “one of the first to really embrace AI innovation at the university. They were willing to test machine learning solutions and see what students were capable of.” One of those students was Joshua Famous ’22, ’23 MEng Eng, a computer science major as an undergrad who didn’t participate in the Challenge but had impressive technical chops and a desire to apply his skills to socially beneficial AI. He joined the Advance program in 2020 and took the lead on the admissions project.

Now in its third year, the project has evolved from a simple idea to a working prototype that is almost ready for real-world trials. The system works by scanning a PDF of a high school transcript and then using AI to extract and sort relevant data based on Penn State requirements. This is a significant technical challenge: It involves not just extracting textual data but understanding its meaning. For example, the AI system needs to be able to tell whether a student took AP classes in high school, when they took those courses, the grade they received, and that grade’s credit value. The challenge is magnified by the variety of fonts and transcript formats submitted by applicants from around the world.

Famous and his team have made impressive progress. They spent last summer integrating the components to demonstrate that it works at scale; the next step is to test it on a broader dataset of real transcripts, then work with the admissions office to develop a plan for implementation. If it works, the system will automatically sort the transcripts for admissions officers to review. “This summer was hopefully the last piece of the puzzle,” Famous says.

Coudriet says the success of the admissions project has attracted the attention of administrators across higher education—he’s had “preliminary discussions” with other Big Ten schools interested in using Nittany AI’s tool for their own admissions process. “The basic premise is, AI plus data enables new, more efficient processes,” Coudriet says. “Universities hold the key to successful AI solution development—data. If we can combine our data and various AI tools to help solve problems for Penn State, why not extend these solutions to other universities and create an innovation engine that can generate revenue for Penn State?”


Many of the students who have participated in the Challenge and Advance programs have been surprised to learn that it wasn’t necessarily the technical skills they acquired that have proven most useful in the job market. Rather, it was the soft skills, such as client communications, team building, and project management, that they say benefited them most when they entered the workforce. “My experience with Nittany AI was literally all I talked about in my interviews,” says Elisii. “It showed I was able to manage a team, design a project from start to finish, implement it, and then pitch it. Recruiters love hearing that type of stuff.”

Warren, who was hired as a user experience (UX) designer at KiwiCo, an edtech company, concurs. “I don’t know if I would’ve gotten into IT UX as a career if it hadn’t been for the Challenge,” she says, “[but] I was infinitely more prepared than I would have been if I hadn’t done it.”

Allen Puy is a manager of mission and systems engineering at Lockheed Martin Space, where a big part of his job is helping the organization recruit top talent. For 25 years, he has been one of Lockheed Martin’s liaisons to Penn State, helping the company identify talented graduates via job fairs and student meetups. He says the benefit of recruiting alumni of Nittany AI is that “they’re actually hiring students and building a team based on what we’re interested in,” says Puy ’86 Sci. “They select students who meet the profiles we would be looking for, and we get the chance to interact with them in really rich ways so we can understand their capabilities, their aptitudes, and how they approach problem-solving. It’s a model that was really attractive to us.”

Lockheed Martin has hired three students through its partnership with Nittany AI, and Thomas Foltz, a senior majoring in computer science, spent his summer interning in their space division. It’s an opportunity he credits to his participation in the Nittany AI Challenge and his work on the SEPTA project. “It was really beneficial, because we were able to have a tangible project that helps customers,” Foltz says. “It’s rare for a student to have the experience of being able to take an idea and build it into a proof of concept.”

The post-grad success of alums such as Elisii and Warren is validation of Nittany AI’s mission, showing how the program’s impact ripples well beyond specific projects. The students gain both hard and soft skills that will be increasingly valuable in a rapidly changing world, and at a time when so many are anxious about AI’s impact on everything from employment to warfare, Nittany AI is demonstrating how these new technologies can be a force for good.

“Being an innovator means taking risks,” Coudriet says, “and Penn State is allowing us to do that.”


Daniel Oberhaus is the author of The Silicon Shrink, a book about AI and psychiatry due to be published in fall 2024.