While most of the debate over artificial-intelligence technology at colleges centers on the classroom, there’s another hotbed of AI adoption on many campuses — the marketing and communications office.
It’s not hard to imagine why. While colleges (outside of athletic departments) took decades to warm to marketing and branding — sometimes tarring the practices as unwelcome intrusions from the world of business — college leaders now readily turn to marketers to broadcast their institutions’ distinctiveness among their peers, to help them compete for the dwindling number of traditional-age students, and to bolster their images with stakeholders. That has meant more work for a growing number of college marketers — Website content! Press releases! Photos! Social-media posts! Videos! A brand new social-media platform that needs yet more posts! — but not always commensurate personnel or dollars. “Ask a [college] marketer if the budget is big enough,” says Andrew Sogn, marketing and social media coordinator for South Dakota State University, in Brookings. The reliable answer is a resigned “no.”
So when new technology comes along that promises to whip up ideas, write copy, fix photos and potentially even create video, how can you resist at least checking it out? It’s like hiring an intern who works forever. You never even have to buy it pizza.
Except it’s not like that at all, says Bart Caylor, a consultant who works with colleges, the co-host of The Higher Ed Marketer podcast, and an unabashed AI evangelist. “Because they’re not an intern that doesn’t know anything,” he says. “They are a world-class scientist, and they just need to understand the way you do things.”
While many college marketers are excited about the possibilities of utilizing AI and how it can help them do their jobs better, some are also concerned about some of AI’s growing pains — its biases, its fabulations, its appropriations. Some worry that it may cheapen their work or even cost them their jobs.
One big reason it’s hard for college marketers to make headway with the people and resources they already have is what they refer to as the “copy shop” mentality. They are forever being asked to come up with text for this release or that post, or to make improvements to this flier or that slide deck, which often leaves little time for strategic work that can have more substantive effect on institutional priorities.
AI could change that. “It’s kind of like Tony Stark climbing into the Iron Man suit,” Caylor says. “He is still Tony Stark, but the Iron Man suit allows him to have these amazing powers that amplify what he can already do.”
Budgets have increased for college marketers in recent years. Total median marketing and communication budgets for doctoral universities jumped from $2.5 million during the 2019-20 academic year to $3.9 million during 2020-21, a 56-percent increase, according to a survey of chief marketing officers conducted by SimpsonScarborough, a marketing and branding company.
I use it like an editorial assistant or something, another person to bounce ideas off of.
But that amount made up a median of less than 1 percent of total institution budgets, and 68 percent of the marketers surveyed disagreed or strongly disagreed with the sentiment that they had “the budget needed to meet the expectations of leadership.”
College marketers were poised to seize on generative AI, Caylor says. Marketing isn’t a traditional part of academe — it moves faster and changes more, he says, noting the ways that first the internet, then social media, have transformed the industry from the days of viewbooks. Most marketers aren’t native to a higher-ed environment, joining a college staff after at least some experience with a nonprofit or in business, where approaches, and the pace of innovation, may be different.
At the American Marketing Association’s Symposium for the Marketing of Higher Education last fall in Chicago, every session devoted to AI filled up early, and conversations buzzed with questions and tips about the technology.
For Don Hunt, senior vice provost for enrollment management and services at North Carolina State University, in Raleigh, deploying chatbots and other earlier forms of machine learning was about looking “for low-hanging fruit that may give some relief to the staff and also improve customer service over all,” he says.
“Everyone has their different comfort levels with how we’re leveraging AI,” he adds, but if more sophisticated generative-AI software can further increase efficiency and improve the student experience, “it’s not a hard conversation for the division to embrace it. It’s just how we embrace it.”
It seems impossible, but this generation of generative-AI platforms is still brand new — ChatGPT, the most avidly discussed software among college marketers, is barely a year old. In the span of a little more than a year, however, ChatGPT and the different AI options that have erupted in the space around it have given marketers a sandbox full of new tools and glimpses of how they might revolutionize their work.
To date, college marketing and branding pros are exploring generative AI in three main areas: idea generation and brainstorming, content generation, and working with images.
Marketing and branding are creative endeavors, requiring marketers to find new ways to use their messaging to connect with their audiences. That can mean racking your brain to come up with a tagline to get across an institution’s mission or appeal in just a few words, social-media copy that will snare the fleeting attention of a 16-year-old, or, sometimes, an ad that might resonate with a 37-year-old salesman enough to get him to commit to an MBA program.
By feeding parameters about a certain type of message, audience, and tone into software such as ChatGPT, a marketer can generate dozens of suggestions in seconds.
“I use it like an editorial assistant or something, another person to bounce ideas off of,” says Sogn, of South Dakota State. “It’s certainly taking some of that weight off of the writer’s-block moments.”
Most widely available generative AI is trained on the broadest possible spectrum of data, so general questions directed to it don’t necessarily return keen insights.
Sarah Bjorkman, who spent more than a decade working in communications and marketing at the University of Minnesota-Twin Cities, most recently as the director of communications for the School of Public Health, says she sometimes turns to AI for inspiration or ideas, and “a lot of what we get back from those idea-generation or brainstorming examples isn’t knock-your-socks-off,” she says. “But it always produces insights that we hadn’t thought about — ‘Oh, yeah, we should incorporate this into our campaign.’”
All of the marketers contacted for this story who are using generative AI in their work say they only use written material it produces after first running it through one or more staff members — none are willing to trust their institution’s messaging to a machine just yet. But several of them are using AI to help them generate and wrangle the reams of text that many marketing shops are expected to produce or oversee, and, they say, it promises to be a game changer.
Hunt, the North Carolina State vice provost, sees a potential future for AI in helping with the kind of granular marketing that can make the difference with students in a saturated media environment, where people are awash in content and can tune out the less clever and persistent.
Hunt is in the process of trying to better market some of the university’s programs, such as the College of Natural Resources.
“Think of how Amazon engages you about their products, getting you to follow breadcrumbs to potentially purchase something,” he says. “So you get a message here, and then you may get an email there, or you may see a popup.”
Such efforts, known as “drip campaigns,” target specific consumers at specific times with messages based on their actions at each step and their progress toward an actual investment.
They can be labor intensive for marketers — involving, for example, more than 50 different messages, Hunt says — and the process for designing the campaign, and crafting that many messages, could take a staff member several months.
Now, generative AI “could write all the content for you, and you can review it,” Hunt says, and a drip campaign for a program “could possibly be launched in a couple of weeks.”
Even though marketers are still reluctant to use AI-generated words without oversight, they’re eager to have them help with the sheer volume of words they deal with.
Sogn has used ChatGPT to shorten speeches for campus leaders, asking it to lop a few hundred words off a draft, or to help him generate captions for social-media videos when inspiration starts to run dry. “It’s never a direct copy-paste,” he says, “but a lot of times, it’ll spark an idea or it’ll have a phrase in there that I hadn’t thought of yet.”
One of the most immediately effective uses of generative AI for marketers comes in manipulating images — the ability to quickly alter photographs that might previously have been unusable or required hours of painstaking editing to fit the bill.
Over video from his office at Emory University, Luca Magnanini, the director of enrollment marketing and communications, shares a promising recruiting image: students walking across the institution’s Atlanta campus amid autumnal colors while wearing shorts — a subtle lure for students sick of dour Northeastern weather.
There’s just one problem. The blurred yellow line of a ribbon of caution tape cuts across the bottom quadrant.
In the old days Magnanini would have needed a staff designer to spend an hour or so using software tools to remove the tape from the image bit by bit. Now, he loads the image into the Generative Fill function in Photoshop, types in a few commands, and the generative AI now built into the software erases all but an easily cropped trace of the end of the yellow tape in a few seconds.
“That’s super easy, right?,” he says. “And that’s just a quick example.”
It may sound like a small thing, and it is, but it solves a longstanding problem plaguing marketers: the almost-perfect image.
College marketers cite many examples. There’s a great photograph of nursing students, but they’re wearing jewelry, which is not allowed in professional settings. A summer scene that would be more useful if it were taken in the fall. A fantastic photo of donors at a gala, but some of them have drinks in their hands.
Generative AI has the potential to quickly and efficiently remove jewelry, give summer foliage autumnal tones, or turn cocktails into waters all around.
Not everyone in college marketing and branding is sanguine about generative AI and the sector’s early embrace of it.
Before she founded Campus Sonar, a company that consults colleges, Liz Gross was a college marketing and communications professional herself. She would also like to see more college marketing offices get away from the campus copy-shop model, but she worries that AI will only exacerbate it. “I really fear that generative AI is going to be used just mostly to create more output,” she says, “and feed into that high-output, publicity-only model rather than better strategies.”
She, and others, are also concerned about colleges trying to tailor marketing messages using the wide net of generative-AI source data. Software such as ChatGPT is, by design, trained on as much data as possible. While that can bring advantages in terms of unexplored angles or information, “I don’t understand why you would want to start with the generic and then try to somehow turn it into your own message,” Gross says.
In an era when colleges are already working hard to differentiate themselves from their competition, ceding any part of their identity to generative AI may be a mistake, says Teresa Valerio Parrot, a principal of TVP Communications, a company that consults colleges.
“If you’re blendering other people’s ideas into a smoothie to become your own, does it really have the same weight and influence and clarity that you hope it has?” she says.
She worries there could even be larger ramifications to the use of AI for college messaging. “At a time when higher education is being questioned for its value, its authenticity, and, in many circles, its truthfulness,” she says, “we need to be thinking about whether or not the tools that we can use to get to an endpoint more quickly are reinforcing some of the concerns that we’re hearing from the public about how we do business.”
There are also more concrete hazards attached to the use of AI for colleges to consider, says Gil Appel, an assistant professor of marketing at George Washington University.
Generative-AI software has, so far, often shown itself prone to racism and sexism, he says, due to the nature of the data it’s fed. When asked for images of a productive person, for example, some AI software may return images of light-skinned younger men.
One of the most sacred and holy of expectations in higher education is that we aren’t plagiarizing, and we also are being authentic.
Using generative AI to create images could be especially problematic for colleges, Appel says. Just as text-based generative-AI software is trained on text, image-based software is trained on images. While there is only now beginning to be scrutiny of the use of copyrighted written material by software companies, artists have already pushed back on their work being subsumed and appropriated by AI software. For example, a group of illustrators have filed a class-action lawsuit against several AI companies over using their work in training software.
Appel says that few AI platforms can generate copyright-protected images, or compensate the artists for the work the software was trained on. If a college used an AI-generated image not protected by copyright, Appel says, it could expose the institution to legal action.
Gross gets that college marketers are loath to miss out on cool new tools that might help them do their jobs more efficiently, but she thinks college marketers and leadership should spend more time reflecting on AI and what it means.
“Institutions need to understand what generative AI is, why they would value it, and be able to have a nuanced and informed conversation with any of their partners or vendors,” she says.
Businesses are now offering AI as a value added to all kinds of services, and that may provide benefits, but “when you start to integrate it into your business and organizational practices, there needs to be a stronger understanding of what it means from a risk perspective and from a brand perspective.”
The Public Relations Society of America has released a set of ethical guidelines regarding the use of AI. But perhaps the first rule of AI for college marketers should be do no harm to the brand, and that means protecting a sense of authenticity above all, Parrot says.
“One of the most sacred and holy of expectations in higher education is that we aren’t plagiarizing, and we also are being authentic,” she says.
If a stakeholder feels they’re hearing from a machine, rather than a person who represents the college and has their interests at heart, the consequences for that relationship could be dire.
Last year, Vanderbilt University made headlines as one of the earliest higher-ed adopters of ChatGPT when the university’s office of equity, diversity, and inclusion used the software to write an email to students in response to a shooting at Michigan State University. Vanderbilt officials later apologized and called the use of ChatGPT “poor judgment.”
That’s one of the reasons Hunt, the North Carolina State vice provost, is exploring generative AI cautiously. “I don’t think you can just blindly go down the road and just adopt it without serious consideration,” he says. To do otherwise, “can set you up for a very disastrous outcome.”
Caylor, the consultant, has been around higher education long enough to have built his first college website in 1994. Back then, many leaders he spoke to were dubious whether they would make any difference. “They were like, ‘Yeah, Bart, that’s good,’” he says, “‘but I really can’t imagine anybody surfing the web to come to our school.’”
Caylor thinks that, while it’s still early days for generative AI, it will have a similarly seismic effect as the internet on nearly everything, including college marketing.
It’s a tool that almost everyone will eventually adopt, he adds, the way almost everyone now writes with a keyboard rather than with a pen or pencil and paper. (And many use AI-assisted software like Grammarly without a second thought.)
Part of improving the efficacy of AI lies in improving the written prompts used to elicit responses. The more specific and tailored a prompt is, the more useful the outcome is likely to be.
If you ask ChatGPT to write an email to employees about an engagement survey, “it’s going to spit out a very generic or jargony sort of messaging,” says Bjorkman. “Or you could say something to the effect of, craft a message from the dean to the School of Public Health faculty and staff asking them to participate in the upcoming employee-engagement survey — you’re gonna get a much more targeted and customized message.”
Sogn, of South Dakota State, says he’s seen more marketers listing certificates for writing AI prompts on LinkedIn.
Even the college marketers who aren’t using AI now probably will be soon. It’s already integrated into many common software platforms, such as Adobe products, and that will almost certainly continue.
John Drevs, associate vice president for digital marketing and communication at Loyola University Chicago, is hearing AI as part of the sales pitch from vendors more and more. He’s interested in exploring AI, but if things continue as they are, AI will come to him. “I don’t necessarily trust that, what I’m looking for it to do, it’s going to do as well as either myself or my staff,” he says. If it’s built into the existing technology that he and his team already use, however, “I don’t have to learn it, it’s just going to be integrated into the business process of what I’m doing now. So it’s easy to adopt that.”
Drevs, like other college marketers, has bigger concerns, too — about the way generative AI will affect peoples’ jobs, both in society as a whole, and among he and his peers. “In marketing and communications, so much of what we do is an understanding of people and their motivations,” he says, and “the human element in the creativity we’re offering in our work.”
He loves the idea that AI could help college marketing offices operate more efficiently so that they could focus their work on bigger, more strategic projects, but he worries that the pendulum could swing the other way, “where we don’t need as many people to achieve our objectives,” he says. “Whenever you start having conversations like that, you start seeing faces and names and families behind people’s jobs.”
Sogn believes that college marketers worried about AI taking their jobs are ultimately fretting about the wrong thing. “AI is not going to take your job,” he says. “AI will take the job of the marketer who refuses to use AI.”
Even Hunt, who has moved cautiously on AI so far, feels generative AI is an inevitable part of the future. “The adoption of AI is going to be predicated on the culture of the institution, and what the culture will allow and won’t allow,” he says. “But the market is changing and the student population and their needs are changing so fast, not adopting AI, I think, is going to be a mistake for many enrollment managers and institutions.”
