AI goes systemic in higher education

While academia generates innovative research that leads to industry booms, it’s often the industry that spearheads novel organizational structures. Take the chief artificial intelligence officer (CAIO) and the organizational AI strategy it inspires, for example.

CAIOs first emerged in the private sector around 2023, following suit in government and then higher-ed institutions like colleges and universities. But even with this history, there’s no manual, says Amarda Shehu, CAIO of George Mason University and professor of computer science and associate dean for research.

GMU certainly isn’t the only higher education institution with a CAIO. Others like Canada’s Western University and California’s Sacramento State University have them, too. Yet because of Shehu’s background building programs from the ground up, like the division of information and intelligent systems at the National Science Foundation, she understood that a systemic AI infrastructure was needed to propel GMU into AI innovator territory.

The AI approach at GMU is both strategic and generative, whereas in other places, the generative part may be missing, she says, meaning that staff, faculty, students, and third-party providers all work together to shape, build, and iterate solutions that benefit the entire community under a university’s guideline of workforce and economic development.

As part of what Shehu calls AI2Nexus, a model for universities to drive responsible AI alongside societal impact, GMU seeks to both integrate AI at all levels and generate new AI use cases. For example, they launched a firewalled platform called PatriotAI this summer, which includes AI tools like a chatbot and agents for document analysis, exam preparation, food access resources, syllabus refinement and more. While some tools have access blocks, the platform doesn’t exclude any population from the university, but rather acts as a sandbox where people from different groups can use and even create tools.

“The faculty can upload their course materials, and late at night when faculty isn’t checking emails, the student can decide this is when I’m going to level up, engage with the material, ask questions, and maybe create mock-up quizzes and tests,” she says. “Students are already out there using ChatGPT and others, but we don’t want them to upload material into websites and have potential intellectual property or privacy issues.”

While PatriotAI comes with an initial lineup of agentic solutions, creation is in the cards. “We’re providing the opportunity for members of our community to directly influence what we provide next, and we’re kind of bootstrapping it in house,” she adds.

A team effort

Shehu doesn’t work alone on GMU’s AI ecosystem. Charmaine Madison, who joined as the university’s CIO last year and formerly worked in IT at the CIA and US Air Force, is equally driven to make AI and the cybersecurity that comes with it a cornerstone of the university’s identity.

Madison is focused on maturing GMU’s smart campus vision, which transforms the physical and digital campus into a new kind of living organism. Still at its early stages of maturation, though, Madison intends for it to result in operational efficiencies and cost savings by tracking building occupancy, heating and air conditioning, and accessibility metrics. Noting that GMU has been recognized as a highly accessible campus, if AI and technology can be better leveraged to provide assistance around campus, and facilitate and enhance learning, that’s the primary aim, Madison says.

GMU is classified as an R1 research institution for its high research activity, but it’s not like many higher education institutes that focus on just the research aspect of AI. “We’re also prioritizing ethical and transparent AI practices,” Madison adds. But that doesn’t mean research isn’t a priority. “We want to catalyze new research, new discoveries, and new innovation, whether that’s foundational in building AI technologies, or in AI-enabled advancement of other scientific disciplines, like physics, biology, or bioengineering,” says Shehu.

As part of their overarching AI strategy, GMU endeavors to innovate across teams and organizations. This means tapping into departmental or demographic expertise, but also leaning on partnerships within and beyond the university. Shehu co-chairs the AI in Gov Council alongside Richard Jacik, chief digital officer of government digital transformation firm Brillient Corporation.

By joining government leaders from all levels with industry organizations that design government tech solutions and AI researchers, the broader governmental ecosystem gets to see AI solutions from ideation to prototype and beyond, all of which are tested in the university’s secure environment.

Plus, there’s the Virginia Has Jobs program, a statewide initiative in partnership with Google that seeks to fill gaps in AI education for students and the existing workforce. The program, Shehu says, spotlights GMU’s graduate certificate in responsible AI, and the new master’s in AI, as ways to build skills.

Then there’s the Microsoft partnership, with which GMU recently entered into a five-year contract. “Some say this is too expensive, too difficult, and we don’t know the rules well,” says Madison. “But the governing body we’ve built with our community and the Commonwealth [of Virginia] has allowed us to continue to move forward, and that’s something others have wrestled with. Hopefully, we’ll get more peers, because it would be great to have more collaboration among the institutions in this space.”

The greater university community

Beyond GMU, the higher education space continues to push forward with both individual AI solutions and broader strategic implementations. Under the leadership of CIO Chuck LoCurto, VP and CIO at Bryant University, the school is rolling out AI education for students, faculty, and staff, as well as the tools to make it accessible. “It’s what we teach and how we teach,” he says.

Bryant launched a generative AI chatbot called AskTupper earlier this year and is preparing to launch it publicly on their website without the need to sign in. With the ability to answer questions about policies, resources, and events, the university intends for the tool to cut down on repetitive questions from existing students, prospective families, and more.

The school has also evolved its annual app-a-thon into an AI-powered prompt-a-thon, where students helped ideate and build AI-powered concepts using AI itself. For instance, one winner, ClubMatchAI, is an app that matches students with clubs that fit their interests and personality.

Bryant’s AI technology partner, alliantDigital, further helps build out the school’s ideas, which includes the AI tutor Strategy Guru and marketing chatbot Brand Guru, and more tutors are planned to get rolled out for use across different populations. Staff members get their own AI 101 lessons, too, and have enterprise access to LinkedIn Learning, which has thousands of courses on key AI topics. “We’ve been encouraging the use of that,” says LoCurto. “In fact, I’ve made it a performance review objective that every employee of mine must take two AI courses on LinkedIn Learning, and I suggest they take prompt engineering because it’s the most relevant.”

As for Bryant’s overarching AI strategy, it’s both top-down and bottom-up. Individual departments recommend use cases to improve their workflows, and some of these use cases are currently under review to develop AI solutions. “We’re a school that’s not going to create the next LLM, but we’re going to be known as how best to apply AI,” LoCurto adds.

Whether at GMU, other higher education institutions, or public and private organizations, research shows that deeply ingrained AI strategies are more likely to bring returns. The latest findings from the in-house innovation lab at workplace collaboration software company Asana says those that are successful with scaling AI are 154% more likely to follow a centralized deployment model. As higher education seeks to follow the industry’s example, with the added pillars of research and learning, the institutions that get ahead may be the ones building AI into every part of the enterprise.