Schmidt says the school will emphasize internships and research opportunities to prepare students for AI-driven careers and graduate school programs, while incorporating ethical integration of AI in classes and pursuing strategic partnerships with business and industry.

“Business acumen is going to be important,” he says. “We’re working closely with our partners at the business school to give our CDSP students an opportunity to minor in business.”

A business minor combined with technological skills, internship experience and a liberal arts background is a potent combination, he says. “People who have that combination of experience and skills are going to be a lot more competitive. If you really want to future-proof yourself as a college graduate who will hit the ground running and be marketable right out of the chute, it’s going to be this liberal arts foundation, which is about communication, reading and writing, coupled with technology.”

One of the consequences as AI becomes more pervasive is a reluctance to trust written material because of uncertainty about who wrote it, he says. “Employers increasingly are going to be looking at how people perform, how they present themselves.” To maintain competitiveness, Schmidt envisions continued engagement with W&M graduates as they progress in their careers, through specialized courses and webinars.

As part of W&M’s commitment to be a career partner for life, Schmidt says, “We should be helping to inform and enrich and empower our alumni by giving them access to AI tools that will make them more effective.”

SECURITY AND PRIVACY

“Our information landscape is changing in ways that present new problems and new gaps in our ability to predict what’s coming,” says Trenton W. Ford, an assistant professor of data science at William & Mary and director of the Ford AI and Society Lab. China, for example, used to have clearly defined paths in which to communicate with American citizens, he says. Through social media, “those paths are blown wide open in terms of the ability to engage and interfere and influence.”

Ford, who teaches courses on machine learning and generative AI, was a featured expert in the recent NATO academic conference on societal resilience in the digital age. Held in London from June 30-July 1, the conference considered how democratic societies can confront threats from disinformation and cognitive warfare in the form of AI-generated content and deepfakes.

Ford spoke about how technological tools such as those being created in his lab can be used to combat information warfare. He is the principal investigator on two projects with the Defense Advanced Research Projects Agency (DARPA), the central research and development arm of the U.S. Department of Defense. The projects simulate human behavior using digital “twins” based on the demographics of a community and leverage generative AI to create a modeling and simulation framework. The idea is to predict how people in a given community might respond to an event such as a disinformation campaign by an adversary and develop ways to defend against that.

The more sophisticated large language models like ChatGPT become, the more difficult it is to detect whether a message is coming from a chatbot, Ford says. For example, “it is entirely possible that in some of the most fervent arguments being had on social media, one of the parties is not human and the other party does not know.”

Experts at the conference considered how democratic societies can maintain trust in their institutions in such an environment.

“The large takeaway from the conference was, if you can’t quiet the messaging that you think is giving a distorted view of reality, you have to tell your story more convincingly,” he says. Appealing to the shared values of a country’s population can help guard against information warfare: “People will get better at digital literacy if they see it as patriotic to not be easily persuaded by enemies.”

It’s a path he discovered as an undergraduate when he came across an application to work on a project to monitor populations of sandhill cranes in central Nebraska by using AI to analyze thermal imagery. Part of the Institute for Integrative Conservation’s Conservation Research Program, the project involved William & Mary’s Center for Geospatial Analysis, the U.S. Fish & Wildlife Service and the U.S. Geological Survey.

Although he had not taken a biology class since ninth grade, Luz-Ricca was drawn to the idea of applying technology to an environmental challenge. The Fish & Wildlife Service (FWS) was looking for a more efficient way to count sandhill cranes, which are being monitored because overhunting has diminished the birds’ numbers in the past.

Having a person manually count the cranes would be time-intensive and not cost-effective, Luz-Ricca says. It’s also difficult to count the birds during the day because they’re dispersed over a wide area. They roost at night, when they’re harder to see, and there could be up to 4,000 or 5,000 cranes clustered together. That’s where thermal imaging comes in, since the birds’ heads give off heat.

“It was an interesting methodological problem, too, because it we couldn’t just take a machine-learning model off the shelf and apply it,” he says.

As a rising junior, Luz-Ricca worked with the conservation partners to develop and train an AI counting model, test which altitudes work best for the planes, collect the thermal images and figure out how to adjust for false positives — so the model doesn’t mistakenly count rocks as birds, for example. He then became the lead author of an article about his research in the journal Remote Sensing in Ecology and Conservation.

Now in graduate school, he is researching the global effects of hunting in pushing mammals toward extinction.

“Artificial intelligence tools help us take raw field studies that only look at one population in one place and bring these together into a large aggregate data set,” he says. “If we bring enough of them together, we can start to make generalizations based on global patterns.”

He uses AI in his personal life, too, as the father of children ages 9 and 13. “Just last night we were researching jet skis — should we get them and which one should we get, and all the criteria that we wanted as part of those considerations,” he says. “We were using AI to game out these different scenarios and help us with decision making and total cost of ownership. It was actually a very useful way to utilize that technology.”

Wallig visits William & Mary’s campus regularly as a member of the academic advisory board for the MSBA program, and he participated in the business school’s “AI in Business Education Summit” in April. At the summit, faculty members described how they’re using AI in their classrooms. In one memorable example, David Long, the J. Edward Zollinger Associate Professor of Business, had trained an AI model to impersonate the CEO of Boeing as part of a case study.

“Students could interact with that persona and ask questions like, ‘Why did you make this decision?’” Wallig says. “And I thought, ‘That’s such a great extension of the classroom that doesn’t detract from the professor’s authority or ability to teach or communicate their message. Students can do this on their own and then engage in a different level of dialogue in class.’”

In addition to a net job gain, Wallig sees an opportunity for people to do more interesting and fulfilling work as a result of AI.

“We are entering an era of more entrepreneurship and less 9-to-5 hierarchical work in the same place for a long period of time,” he says. “So if you’re entrepreneurial and flexible and creative, and you’re a William & Mary T-shaped person — with broad understanding and deep expertise in certain areas — I think you’re going to be great.”

Turville agrees: “You have to focus on functions and topics where you are becoming the expert and they are not things that can be done by AI — conceptually based, thoughtful work.” In the Professional Development Week talk, she advised listeners to keep up to speed by taking AI courses, listening to podcasts and learning how to leverage new technology.

At Accenture, her team uses generative AI tools to gather and sort through volumes of federal government data to research statutory requirements or to find out, for example, what policies a specific benefits program may have paid out in a given year and how those dollars compared to previous years.

“Organizations have massive documents explaining policies and procedures to obtain a benefit or complete an action like applying for a patent,” Turville says. “We load all of these public documents into AI tools, reinventing the process, cutting out wasteful or non-value-added steps.”

Like Wallig, she is optimistic about the potential for AI tools to reduce mundane tasks. She also believes that with greater efficiency, government agencies will be able to process requests for loans or benefits more quickly.

“Sometimes you hear about people applying for programs, and then it takes months to get approved,” she says. “But what if you were able to apply and the person who has to make that decision was able to access all the necessary information in about two minutes, and then could make that decision right away? Those are the things that get me super excited.”