Artificial Intelligence

Dive Brief:

• A large majority of college students — 90% — report using artificial intelligence in class at least occasionally, according to new survey data from ed tech provider Instructure.
• Nearly two-thirds of college instructors, 61%, said the same. But the survey, released Tuesday, also found that only 11% of higher ed instructors received comprehensive AI training, while 41% said they had not received any formal training on AI.
• Educators are looking for support in the form of hands-on training, clear guidance from their institutions about AI use, and ongoing professional development on the topic, according to Instructure.

Dive Insight:

Higher education has been grappling with AI use in college classrooms and boardrooms since late 2022 when ChatGPT first launched. Much of the notoriously staid — and, at times, slow-moving — sector has been left to play catch-up in the four years since, as the technology has progressed in staggering leaps and bounds.

Some high-profile institutions, such as Arizona State University and the California State University system, have struck deals with AI companies to grant students and employees access to their tools. But many colleges are still developing AI use policies, if they have started the process at all. 

Either way, AI is already “part of how students learn and educators work,” Melissa Loble, Instructure's chief learning officer, said in a Tuesday statement.

"The challenge now is making sure people have the training, judgment and clear expectations to use it well,” she said. “That requires practical support for educators and thoughtful boundaries that keep critical thinking, human judgment and meaningful learning at the center.”

The Instructure survey polled over 1,100 higher education and K-12 educators, college students, and parents of K-12 students. College students reported being the most active users of AI and the most likely to see an upside to the technology, with 94% identifying at least one reason to be optimistic about AI’s educational applications. A majority of surveyed educators also reported at least one reason for optimism. 

But among both college educators and students, 65% expressed concern that AI can sound confident when it is wrong. They also reported worries about a potential overreliance on the technology and loss of critical thinking and deep learning.

The new survey offers some insight into what is expected of educational institutions. 

Every demographic surveyed by Instructure said they expected schools to teach the ethics and responsible use of AI, as well as literacy for the technology.

Respondents also showed more support for the use of AI in support tasks, such as finding resources, compared to using it for academic decisions or grading assignments.

Other polling has found increasing skepticism around AI. A Gallup survey conducted in December found increasingly negative sentiments among Gen Z adults about AI. Some 83% called it at least “somewhat” likely that AI designed to complete tasks quicker “will make learning more difficult in the future.”

Students have also voiced their displeasure with AI in more direct ways. Newly minted graduates at the University of Arizona booed former Google CEO Eric Schmidt when he praised AI during a commencement address in May. At least two other spring commencement speakers at other colleges faced similar opposition. The same month, Columbia University students joined faculty in protesting the use of AI to read graduates' names during commencement

 There are people who fear that artificial intelligence will render human beings irrelevant in the workforce. Denise Kleinrichert is not one of those people.

A management professor at San Francisco State University, Kleinrichert predicts that the use of AI will become as common as the use of cell phones, and that organizational departments to oversee AI’s use will become as ordinary as human resources departments. 

“Is it going to completely replace all human beings? I can’t foresee that in our lifetime. Or a future lifetime,” she said. “It’s just changing the way we do business.”

But for that rosy future to become reality, Kleinrichert said, employees must learn how AI works, where it is biased, when it might threaten privacy and how to recognize when an AI agent’s output is just plain wrong. 

That’s why she is teaching a course in AI ethics and compliance as part of San Francisco State’s graduate certificate in ethical artificial intelligence, which has been offered since 2019. “Employers are looking for employees who have some savviness about what is ethical and what isn't when it comes to AI,” she said. “We need to prepare our students.”

This is driving the popularity of courses, certificates and master’s programs focused on AI ethics. Some are designed for students with little or no computer science background. Others focus on how to use AI in a specific field. But at the core of each program is an emphasis on avoiding harm.

“AI concerns everybody,” said Sonja Schmer-Galunder, an AI and ethics professor at the University of Florida. “We need to provide a more holistic education that is focusing on how we can do this safely and ethically.”

What’s an AI ethicist?

A 2025 report from the labor market analysis organization Lightcast found that job postings requiring generative AI skills for nontechnical roles including healthcare, hospitality, education and finance grew ninefold between 2022 and 2024.  And a growing number of jobs specifically require expertise in AI ethics — meaning whether the technology is being used responsibly and fairly within an organization or business.

More than 100,000 jobs for employees with expertise in AI ethics are now posted each year, according to a 2025 study. And while many technical skills can be learned on the job, AI ethics is highly interdisciplinary — touching on data science, business and philosophy.

Anyone using AI must consider the risks versus the benefits of relying on the tool. But for some, assessing an AI tool’s ethical reliability is the core of their job. 

AI ethicists are tasked with making sure that their employer’s use of AI aligns with the law and the organization's own code of conduct. Ethicists combine knowledge about how AI works with basic ethical principles, and they are often charged with running AI tools through trustworthiness audits. 

They can follow best practices for doing these audits, including a federal policy framework focused on cybersecurity and more general guidelines from the Institute of Internal Auditors. Generally, the audits focus on basic questions: How will the risks of using AI be addressed? Is the data used to train the AI models accurate? Are AI tools tested before they are deployed?

Some academic ethics programs even require students to design their own AI tools and then audit them.

“If they develop a medical application, they have to discuss the impact of any errors the system might make,” said Dragutin Petkovic, who teaches the computer science course in San Francisco State’s AI ethics certificate program. “What if it hallucinates? How will that impact the patients?”

Among the questions Petkovic wants his students to ask: Can a human pull the plug on the AI tool in case of a medical emergency? Anticipating a worst-case scenario is part of an ethicist’s job, he said. 

Government regulation adds a statutory layer that ethicists must also monitor. In December, President Donald Trump signed an executive order calling for a federal AI policy. Last month, the White House released a uniform national framework for AI via legislative recommendations to Congress that call for preempting separate state regulations.

Regulating use of the technology is important, said Schmer-Galunder, who teaches an AI ethics course for technology leaders that is part of a master’s degree in AI systems. It helps guard against harm while potentially codifying a society’s goals for how AI can be a benefit. Organizations that use AI, she said, should be deliberate about how AI can help solve society’s problems.

“Whatever we are trying to develop or do or deploy, we should ask whether it is going to lead to human flourishing or clean water or know a positive societal outcome,” she said. This is a challenge in profit-focused corporate settings, she said. 

In her course, however, students discuss ways that regulation could incentivize ethical AI as a competitive advantage. 

Her students work in groups and role-play positions that are common at technology companies: program manager, user experience, budget, technical implementation and ethicist. 

Together, they discuss product development scenarios and the potential risks of releasing AI-integrated products without sufficient testing. Schmer-Galunder, who is developing a separate multidisciplinary master’s degree focused exclusively on AI ethics, cited chatbots as an example.

“We have seen that they can have severe consequences when people are starting to not just anthropomorphize, but have distorted visions of reality,” she said.  When AI chatbots were new, she said, developers discussed how to control aspects of a bot’s personality. But when a 14-year-old in Florida died by suicide in 2024 after allegedly falling in love with a chatbot and isolating himself from others, parent groups blamed the human-like AI bot.  

“I would like to see an increased awareness of that engineering mindset of building something and putting it out there and then seeing what happens,” Schmer-Galunder said, “so there's awareness of being responsible for what you're doing and that you can be the voice of reason.”

The limits of AI

Many academic programs engage philosophy professors to examine AI and poke holes in assumptions about how it works. Even the term “artificial intelligence” is worth questioning, as it can raise expectations about what AI is capable of, said Jay Gupta, a philosophy professor at Northeastern University, in Boston, who teaches an undergraduate course in technology and human values.

“How appropriate is it to read things like autonomy into the performance of AI systems, even as we refer to autonomous vehicles,?” he said, noting that a truly autonomous vehicle could decide to go wherever it wanted. “The language we use can blur, rather than clarify, important distinctions between human cognition and computational processes.” 

AI ethicists, Gupta said, “should retain a philosophical awareness concerning the ambiguity of such terms.” 

It is also true that AI tools make many jobs easier, particularly when web-based research is involved. Matt Cordon teaches a course in AI ethics and law at Baylor University in Waco, Texas. Other courses in Baylor’s certificate series focus on AI in healthcare and business.  

Cordon said that when attorneys write pleadings or memoranda, the manual research can be time-consuming. Using an AI tool to find case citations is efficient — but it also presents big risks. 

“It cites a case, and it gives you a case name and a jurisdiction. But you go to look that case up, and it doesn't exist. It's never existed,” he said. “Lawyers are being sanctioned for copying and pasting these responses into their filings. The profession is being challenged with that problem.” 

An example of this took place last year in California, where an attorney was fined $10,000 for filing an appeal that was filled with fake AI-generated quotes. In September, the state’s Judicial Council issued guidelines requiring court staff to either ban the use of AI or develop their own policy for its use. 

Cordon's students include lawyers, paralegals and others with an interest in how to use AI the most effectively — and without making egregious errors. He said that besides offering false information, such as cases that don’t exist, AI can also create crucial errors related to language. 

Successful legal writing requires the use of precise language, he said, and AI tools just aren’t capable of understanding precision in specific contexts. The words “vague” and “ambiguous,” for example, can be used interchangeably in most situations, he said. But in legal contexts, they have different meanings. A statute that is “vague,” for instance, could be declared void because it doesn’t specify what is illegal. Something that is legally “ambiguous” is simply open to different interpretations.

“Large language models are trained on generally accessible data on the internet, and the internet's not always going to make a distinction between the meaning of ‘vague’ and ‘ambiguous,’” Cordon said. “Artificial intelligence doesn't necessarily recognize all of that nuance. The human has to recognize that. AI won't catch its own mistakes.”

In an increasingly AI-embedded workforce, employees need the skills to think critically about how artificial intelligence is used and whether its output is accurate. 

“Health companies and banks, they need people to be educated. That is the purpose of our certificate,” said San Francisco State’s Petkovic. “Also, people in decision-making roles in state or federal government need some education about how it works.” 

Sponsored content

By Superhuman, from the makers of Grammarly

Most institutions now have an AI strategy. New research shows why so few can see it, measure it, or act on it, and what the institutions pulling ahead do differently.

Nearly every college and university now has an AI strategy on paper. According to EDUCAUSE, 92% of institutions report a work-related AI strategy. By that measure, higher education has caught up with the moment.

The ground tells a different story. Only 13% of institutions measure whether that AI activity returns any value, and among the staff and faculty expected to follow these strategies, just 54% are even aware of the policies meant to guide them. A strategy that nearly everyone has, that almost no one measures, and that half the campus can't see isn't a strategy yet. It's an intention.

“Institutions are not struggling with AI access as much as they are struggling with AI coordination, governance, and measurable integration across workflows,” says Yolanda Wiggins, senior researcher at Superhuman, the company behind Grammarly. The institutions pulling ahead aren't the ones with the most tools or the strictest rules. They're the ones whose AI policy shapes what actually happens on the ground.

Where strategy meets reality

An effective AI policy settles four questions early, before an incident forces the answer: what AI use is acceptable, what data those tools can touch, who is accountable when something goes wrong, and where human judgment has to stay in the loop.

New research from Superhuman and Higher Ed Dive's Studio, based on a survey of more than 150 higher education executives, shows how few institutions have turned those questions into practice. Policies most often cover general-purpose AI tools (87%), but far fewer extend to productivity agents (64%) or academic support tools (59%). Guidance follows the same shape: strong at the institutional level, thinner at the department level, and thinnest at the level of individual faculty and courses, which is exactly where students meet AI. Asked how much of their AI use is actually sanctioned, just 19% of institutions said AI is formally embedded in strategy, policy, and procurement.

That is the difference between an institution that is experimenting with AI and one that is strategic about it.

The blind spot in how institutions buy

The same research surfaces a quieter problem. When leaders evaluate AI tools, they prioritize data privacy and security (62%) and ease of implementation (39%). The criteria most tied to long-term impact rank far lower: scalability across the institution (29%) and alignment with strategic goals (21%). Yet those same leaders name preparing students for the workforce (79%) as the most important outcome they want AI to support. There is a mismatch between what institutions want AI to achieve and what they screen for when they buy it, and it compounds downstream. The biggest barrier to expanding AI use, named by 30% of leaders, is lack of senior management support, which tends to arrive only when someone can show a return.

What it looks like when governance leads

The institutions getting this right aren't the most restrictive. They pair clear, shared expectations with visibility into how AI is actually used, rather than leaning on monitoring alone. At Rowan-Cabarrus Community College, the English division raised transparency and accountability instead of adding restrictions, and cut academic integrity violations from 27 to 1. Rowan-Cabarrus is a Grammarly institution, and Grammarly is now part of Superhuman. Indian River State College built its own AI fluency agent on Superhuman Go, bringing faculty, staff, and students to a shared baseline of AI readiness.

Both are versions of the same move: treating AI policy as operating infrastructure, not a document filed away once a committee signs off.

Closing the gap between AI activity and AI strategy starts with knowing where your institution really sits. The Higher Ed AI Policy Compass helps leaders map their institution's policy posture and find the next move.

For the full research, Superhuman and Higher Ed Dive's studioID surveyed more than 150 higher education leaders on scaling purposeful AI across campus.

Map your posture with the Higher Ed AI Policy Compass

Dive Brief: 

• Gen Z’s negative sentiment toward artificial intelligence has grown over the past year, and many are concerned about it harming their learning, according to a Thursday survey from Gallup, the Walton Family Foundation and GSV Ventures. 
• Anger over AI is increasing among Gen Z at the same time excitement is fading. Nearly one-third of the survey’s respondents, 31%, said AI makes them feel angry, up 9 percentage points from last year. And just 22% said the technology makes them feel excited, down from 36% the prior year. 
• Among K-12 students, 74% said it is “very” or “somewhat” likely that AI designed to complete tasks quicker “will make learning more difficult in the future.” That share was even higher among Gen Z adults, with 83% of respondents sharing that view. 

Dive Insight: 

Gallup’s new survey comes as AI companies are making vast investments into higher education. 

Google announced last year that it would spend $1 billion to provide AI tools and training to colleges and nonprofits. Over 200 colleges have signed up for the Google program so far, ranging from Ivy League institutions like University of Pennsylvania to large public systems such as Texas A&M University. 

Large universities and systems — including Arizona State University, the University of South Carolina and California State University — have also been striking deals with OpenAI, the maker of ChatGPT, to give students and employees access to the company’s tools. 

Against this backdrop, Gallup polled over 1,500 Gen Z students and adults ages 14 to 29 between late February and early March to understand their views on AI. 

The new survey found that AI adoption is slowing among Gen Z. Just over half of Gen Z members, 51%, use AI each week, representing just a 4 percentage point increase over last year. 

However, 52% of Gen Z K-12 students agreed that they will need to know how to use AI if they attend college. 

Still, Gen Z is becoming more skeptical about whether AI will aid their learning. 

Fewer than half of overall respondents, 46%, agreed that AI tools can help them learn faster, down from 53% who said the same last year. 

Part of the growing negative sentiment may stem from concerns about AI’s impact on the workplace, according to Gallup’s report. Nearly half of respondents, 48%, said the potential risks of AI outweigh the benefits in the workplace, a sharp uptick from 37% who said that last year. 

“Gen Z isn’t rejecting AI outright, but they are reassessing its role in their lives. What we’re seeing in the data is a generation that recognizes AI’s utility but is increasingly concerned about its long-term impact on learning, trust and career readiness," Stephanie Marken, senior partner at Gallup, said in a Thursday statement. "Their growing skepticism signals a need for more thoughtful integration of these tools in both school settings and the workplace.”

Concerns about AI have also been mounting in the higher education sector. At Cal State, faculty are urging the system’s chancellor not to renew its $17 million contract with OpenAI, which expires at the end of June. 

In a petition that had garnered over 3,400 signatures as of Thursday afternoon, they argue that ChatGPT Edu, the version built for colleges, is largely the same as the free online version, doesn’t provide accurate answers to student questions, and harms teaching and learning.

SAN DIEGO — Artificial intelligence is rapidly progressing and poised to reshape the workforce in the near future. The higher education sector is in a unique position, as both an employer of millions of workers and a system that prepares students for the labor force.

At the annual ASU+GSV Summit last week, four college leaders talked to Higher Ed Dive to weigh in on two questions: What about AI’s use in higher education are you most excited for? And what has you most concerned?

Editor’s note: Responses have been edited for length and clarity.

Bret Danilowicz, president of Radford University

What I'm most concerned about AI and its evolution is the speed at which faculty will need to redevelop their courses. Faculty traditionally redevelop their classes on a three- to five-year cycle. They're going to have to be updating yearly or semesterly. To keep current with this, we have to change faculty workloads because they can't redevelop the same number of classes at this speed. That would be my biggest concern, because nobody's really talking about it yet.

The greatest opportunity, I think, is for students for whom college is a challenge and they may just be making it across the border with Cs. Traditionally, those students may have had a challenge in getting hired.

If everybody develops AI skills — some of the ethics and some of the basics — that actually elevates the typical student into a place where they are much more employable than they would have been in the past. 

From an access perspective, having students be employed and bring that to support their family or community can make a world of difference. That's that postgraduate success outcome that we're all talking about.

Lisa Marsh Ryerson, president of Southern New Hampshire University

A substantial risk could be giving up decision-making about learning to AI. If your strategy is just buying a bunch of licenses, I don't think that's the best way to go forward. 

I think the strategy has to center on questions like, how do we improve outcomes for all of our learners, and what role will our people continue to play in that work? 

That work is exciting to me. This is a real leadership moment in higher education. 

I'm also excited for my own learning. My executive team and I are going through an AI bootcamp now. One can't even grapple with the questions of how and when to use AI unless we have our own experience and knowledge of the possibilities.

Harrison Keller, president of the University of North Texas

I'm excited about the potential of AI to personalize the learning experience. We could implement at scale the kinds of interventions and provide the kind of feedback that we've aspired to provide for many years. 

Like other major employers, I'm concerned about our ability to upskill and reskill our own workforce rapidly to be able to take advantage of these new technologies. And that includes not only teaching and learning, but how we can leverage AI to improve our operations, to bolster our institutional research capabilities, and to help us be more effective and efficient.

Pradeep Khosla, chancellor of the University of California, San Diego

Most excitingly, it's going to help us significantly improve outcomes without making a significant investment. Every time you want to make significant improvements, you've got to invest a lot of money, right? But I think AI is one of those technologies where some level of investment is going to have a big, broad-based impact.

To balance that, people have to be thoughtful about deploying AI. Think of it as an amplifier. The largest amplification capacity for any human being is an AI agent. You must decide how you want to amplify yourself and what you want to use it for, instead of framing it as having taken over your thinking.

 A long line of students wrapped around Texas A&M University’s academic plaza in early October to receive free training from Google employees on how to use the company’s artificial intelligence tools, such as its chatbot, Gemini, and its research assistant, NotebookLM.  

That same day, about 400 faculty members huddled in a campus building for deeper training from Google on how they could use AI tools to improve teaching and learning in their classrooms and how to effectively and ethically help their students use them as well, said Shonda Gibson, Texas A&M System’s chief transformation officer.

The daylong event was part of Google’s three-year $1 billion initiative to support AI education and job training programs throughout the U.S. The initiative, which launched in August, supports the tech giant’s AI for Education Accelerator that provides higher education students and educators with free access to tools and training and aims to create a community of institutions sharing best practices.

Texas A&M is one of over 200 higher ed institutions that have signed up for Google’s accelerator, according to Lisa Gevelber, founder of Grow with Google, the company’s workforce development campaign.  They include higher ed systems like the University of Texas and University of North Carolina, as well as large institutions like the University of Pennsylvania, University of Michigan and University of Virginia. 

“Every student deserves access to the digital tools and the skills and training to set them up for success. And this is our commitment to supporting that,” said Gevelber. 

The initiative comes as colleges race to ensure their students are prepared to enter a workforce that is becoming increasingly shaped by AI. 

“It’s not just about using the tools,” said Gibson. “We really want our students to have the best experience possible so that they’re fully prepared whenever they leave us to go on and do whatever they’re going to do in their future.”

Professors that integrate AI into their lessons should follow guidance on how to use it to further student learning, said Alexa Joubin, director of George Washington University’s Digital Humanities Institute. 

Without that guidance, students risk using AI as a shortcut by having it summarize information for them instead of actually reading the materials presented and experiencing their lessons, said Joubin. 

Meanwhile, recent research suggests AI could be detrimental to students' skills and outcomes.

A Massachusetts Institute of Technology study released in June found that using AI tools to write essays can impact critical thinking skills and lead to lower cognitive performance. 

Over four months, study participants who used AI tools to write essays underperformed at “neural, linguistic, and behavioral levels” compared with those who didn’t, raising concerns about the long-term educational implications of relying on the technology, the study found. 

Students are essentially “outsourcing key cognitive tasks to AI,” said Joubin. 

The $1 billion initiative

The Texas A&M System joined Google’s initiative, Gibson said, because officials viewed the tech behemoth as the only company offering assistance and guidance at that level. 

Gibson also pointed out the free access to normally paid versions of Google tools, which will be available over the next two years to students attending the system’s 12 institutions. 

Google’s tools can act as a personal tutor for students to help them work through problems and learn material in a customized way, said Gevelber. 

Gemini, for example, has a guided learning feature that can accommodate their learning needs, said Gevelber. The guided learning tool, for example, asks students probing and open-ended questions to spark discussions and dig deeper into the subjects, and it also introduces images, diagrams, videos and interactive quizzes to help them learn topics. 

And NotebookLM allows students to upload lecture notes or class reading materials so it can create personalized quizzes, study guides, summaries and even AI podcasts that present the information in an engaging way, Gevelbar said.

“I’m really blown away by how much Google has offered support for all this work,” said Gibson. “They’ve had a real concentrated, strategic approach to supporting higher education and K-12.”

Texas A&M system students and faculty can also access Google’s Career Certificates, which offers courses on topics such as AI literacy and AI prompting courses that are operated by Coursera

The Career Certificates program is available for free to all students at institutions participating in Google’s accelerator. That program provides training for entry-level jobs in career fields like data entry, cybersecurity, analytics, information technology support and project management, Gevelber said. Since it launched in 2018, more than 1 million students have received certificates, Gevelber said. 

Using AI with ‘strategic intentionality’

The Google initiative is not the Texas A&M System’s first foray into AI initiatives. System leaders have invested in AI training in recent years, in part because they believe the tools help improve grades and lower course withdrawals, said Gibson. 

The system teaches professors how to effectively incorporate AI into their courses through an intensive faculty training effort with the Association of College and University Educators that launched five years ago, said Gibson. Seven of the system’s universities also provide students with AI training via certificates on Coursera’s learning platform through a separate initiative, part of a two-year trial that the system is paying for, she said. 

The university aims to allow students to test and try things, make mistakes, and know how to use different technologies so they are prepared to enter the workforce, Gibson said. To accomplish that, the system needs to train its faculty and staff how to use generative AI tools so they can be deliberate about how they embed them in courses and academic programs, she said. 

“It is our responsibility to teach students to use it ethically and effectively and we have to do that with a lot of strategic intentionality,” said Gibson. 

Professors shouldn’t try to avoid AI, but they should be purposeful about how they allow students to use the tools, said Joubin, who is also an English professor at George Washington. 

In disciplines that require students to pay particularly close attention to word choices, such as humanities, outsourcing cognitive tasks to AI can be particularly problematic. That’s because the words people use to describe the world around them — such as a piece of art — can affect how others perceive it, said Joubin. 

Joubin uses a custom open source AI program in her classes that draws only from her own lectures, publications and materials, she said. It’s also narrowly tailored — the technology isn’t used to help students write essays but rather to help them think about words or topics more deeply, she said. 

If a student asks it a question, for example, the tool will respond with additional questions. She also asks students to rewrite an idea or a sentence in as many different voices as they can, such as from a lawyer, someone at a corporation or a poet. Then she asks the students to use AI to carry out the same task — giving them even more examples of voices.

“We don’t use AI for any task that requires accuracy, such as information retrieval," Joubin said, adding that they instead use it as a "thinking partner."

In other disciplines, like computer science, students could use AI to write code, said Joubin. AI tools, such as GitHub Copilot, can assist people who are coding. 

But asking AI to write code from scratch is “an utter waste of time,” and not a substitute for students building domain expertise themselves, Joubin said. The quality of the AI-generated code is often lacking — and it may actually take longer and use more resources to create, she said. 

“There are many ways to code something, to execute a task,” said Joubin. “For people who think that STEM needs AI, will just do well with AI without this kind of critical thinking or reflection, they are being deceived.”

Google’s guided learning tools, meanwhile, aim to help people challenge their thinking — giving them different points of view and expanding their thinking about a particular problem, said Gevelber. 

At Texas A&M, professors will adapt their AI policies to different situations — ranging from banning students from using the tools in their classrooms, allowing students to use them cautiously under their guidance, and giving students full access to them after some basic instruction, she said. 

“That is going to differ depending on the type of course, the level of the course, what it is that you're trying to accomplish with your students,” said Gibson. But “Gen AI is something we have to be open and transparent about. How do we want our students to use it?”   

SAN DIEGO — Though the long-term effects of artificial intelligence on the job market remain to be seen, the rapidly developing technology has already begun changing how employers discuss work and hiring. Colleges now face the daunting task of preparing students for occupations that could look radically different in a relatively short timeframe.

A 2025 report from the Georgetown University Center on Education and the Workforce found that the rapid implementation of AI could simultaneously ease and worsen skills shortages, reducing the need for workers to complete some routine cognitive tasks but prompting demand for advanced technical and adaptive skills.

College leaders and higher education experts at the annual ASU+GSV Summit this week shared their experiences on teaching students AI-resilient skills and partnering with businesses to prepare them for the next generation of workforce needs.

Offering students resilient workforce skill sets

One of higher ed's most repeated concerns about AI’s evolving abilities is how to prepare students for a workforce that could look significantly different by the time they complete their degrees.

Jeff Strohl, director of Georgetown CEW, expressed doubt that quicker programs for students, such as AI certificates, could be a long-term solution for workforce preparation.

"I'm a little skeptical, because I fear that the things in the labor market are going to change more quickly than teen romance," he said, eliciting laughter during a Tuesday session". 

Instead, Strohl encouraged higher ed officials to consider how AI and skills-focused hiring "are changing occupations not at the occupation level, but at the task level."

Colleges can view jobs as a portfolio of tasks and curriculum as a portfolio of skills, he said. From there, they can teach students skills to accomplish those tasks in any situation.

Fisk University, a historically Black institution in Nashville, Tennessee, is seeking to address this by arming its students with a combination of durable skills — often referred to as soft skills — and technical skills, according to President Agenia Clark.

"One without the other produces a worker, but together they produce a leader," she said during a Tuesday session. "Leaders are what this moment is requiring."

Sukhwant Jhaj, vice provost for academic innovation and student achievement at Arizona State University, said that work experience must be woven through a college's curriculum, rather than being incorporated only at the end of a student's education.

Pulling from his architecture background, Jhaj told attendees that the biggest challenge in the design process is not running out of creative ideas — it's focusing effort on solving the wrong problem.

To that end, colleges must face the changing workforce reality head-on, rather than attempting to innovate for a world that no longer exists, he said.

Working with employers

Early in his higher ed career, Gordon Jones, now president of the College of Western Idaho, hit roadblocks when trying to establish student-focused business partnerships.

In one instance, a multistate equipment distributor ended its workforce development partnership with his institution over reliability concerns and decided to instead develop its own training program, he said.

That instance and others like it illustrate why it’s difficult for higher ed to "develop deep trust with employers, reliably deliver results and ideally coalesce institutional commitment to the needs of these local employers," Jones said.

The College of Western Idaho, a public two-year institution, now has multiple public-private partnerships built on deep-seated collaboration. In one example, two electric vehicle manufacturers taught the college's faculty the entire process for making their cars. Those instructors then created a vehicle technician training program, one of the first for EVs in the region, according to Jones.

This type of partnership can be especially mutually beneficial in rural communities, where businesses tend to be smaller and lack robust resources, according to Radford University President Bret Danilowicz.

The public institution in Virginia has partnered with local businesses to give its students experience solving real world problems. In small-scale consulting projects, students are hired to work about 30 hours on companies' unsolved needs or pain points.

"The business project may or may not involve AI, but our students are using AI in developing the solution for those companies," Danilowicz said. He added that grades and retention rates improved among Radford students who participated in the program.

University faculty also incorporate local businesses' operational challenges into their courses as case studies. From there, students learn to solve them with both traditional and "AI-supplemented business methods," he said.

If colleges are looking to place students in local internships, simplifying how they get credit for that work — and lightening the administrative load for companies — is a crucial way to get buy-in from employers, according to Jillian Low, chief strategy officer at Virtual Internships, an ed tech startup that uses AI to match students with internships.

This could be especially important for a smaller company that "does not have an HR team, doesn't have a campus recruiting team and, most assuredly, does not want to fill out 10 forms across four formats," she said. 

COLUMBUS, OHIO — Artificial intelligence-based products and software for college admissions and operations are proliferating in the higher education world. 

How to choose from among them? Well, leaders can start by identifying a problem that is actually in need of an AI solution. 

That is one of the core pieces of advice from a panel on deploying AI technology responsibly in college administration at the National Association for College Admission Counseling’s conference in September.

Jasmine Solomon, senior associate director of systems operations at New York University, described a “flooded marketplace” of AI products advertised for a range of higher ed functions, from tutoring systems to retention analytics to admissions chatbots. 

“Define what your AI use case is, and then find the purpose-built tool for that,” Solomon said. “If you're using a general AI model or AI tool for an unintended purpose, your result is going to be poor.” 

Asking why before you buy

It’s also worth considering whether AI is the right tool. 

“How does AI solve this problem better? Because maybe your team or the tools that you already have can solve this problem,” Solomon said. “Maybe you don't need an AI tool for this.”

Experts on the panel pointed out that administrators also need to think about who will use the tool, the potential privacy pitfalls of it, and its actual quality. 

As Solomon put it, “Those built-in AI features — are they real? Are they on a future-release schedule, or is it here now? And if it’s here now, is it ready for prime time or is it ‘here now, and we’re beta testing.’” 

Other considerations in deploying AI include those related to ethics, compliance and employee contracts.

Institutions need to be mindful of workflows, staff roles, data storage, privacy and AI stipulations in collective bargaining contracts, said Becky Mulholland, director of first-year admission and operations at the University of Rhode Island. 

“For those who are considering this, please, please, please make sure you're familiar with those aspects,” Mulholland said. “We've seen this not go well in some other spaces.”

On top of all that is the environmental impact of AI. One estimate found that AI-based search engines can use as much as 30 times more energy than traditional search. The technology also uses vast amounts of water to cool data centers.

Panelists had few definitive answers for resolving AI’s environmental problems at the institutional level. 

“There's going to be a space for science to find some better solutions,” Mulholland said. “We're not there right now.” 

Solomon pointed to the pervasiveness of AI tools already embedded in much of our digital technology and argued untrained use could worsen the environmental impact. 

“If they're prompting [AI] 10, 20 times just to get the answer they want, they've used far more energy than if they understood prompt engineering,” Solomon said. 

Transparency is also important. At NYU, Solomon said the university was careful to ensure prospective students knew they were talking with AI when interacting with its chatbot — so much so that they named the tool “NYUAdmissionsBot” to make its virtual nature as explicit as possible. 

“We wanted to inform them every step of the way that you were talking to AI when you were using this chatbot,” Solomon said. 

‘You need time to test it’

After all the big questions are asked and answered, and an AI solution chosen, institutions still have the not-so-small task of rolling the technology out in a way that is effective in both the short and long term. 

The rollout of NYU’s chatbot in spring 2024 took “many, many months,” according to Solomon. “If a vendor tells you, ‘We will be up in a week,’ multiply that by like a factor of 10. You need time to test it.” The extra time can ensure a feature is actually ready when it’s unveiled for use. 

The upside to all that time and effort for something like an admissions chatbot, Solomon noted, is that the AI feature can be available around-the-clock to answer inquiries, and it can quickly address the most commonly asked questions that would normally be flooding the inboxes of admissions staff. 

But even after a successful initial rollout of an AI tool or feature, operations staff aren’t done. 

Solomon described a continuous cycle of developing key metrics of success, running controlled experiments with an AI product and carefully examining data from AI use, including by having a human looking over the shoulder of the robots. In NYU’s case, this included looking at responses the chatbot gave to inquiries from prospective students.

“AI is evolving rapidly. So every six months, you really do want to test again, because it will be different,” Solomon said. “We did find that as we moved forward, we could decrease the number of hard-coded responses and rely more on the generative. And that was because the AI got better, but also because our knowledge got better.”

Solomon recommended regular error checks and performance audits and warned against overreliance on AI. 

“AI is not a rotisserie. You don't set it and forget it. It will burn it down,” she said. “It's changing too fast.”

SAN DIEGO — Colleges are increasingly looking to recruit adult learners as a means of bolstering enrollment amid an anticipated decline in traditional-aged students. In turn, higher education officials are grappling with how to best serve students who have likely not been in a formal educational environment in years.

For some, the answer is artificial intelligence — a topic that continues to divide educators. Detractors warn AI could diminish the value of higher ed by eroding students’ critical thinking skills and leading to an environment where AI both writes and grades classwork. But supporters and ed tech companies tout rapidly advancing automation as a cost-effective solution to some of the sector's toughest challenges.

This week, leaders from two massive colleges that have long specialized in nontraditional students — Southern New Hampshire University and the University of Phoenix — spoke at the annual ASU+GSV Summit and shared insights from their experiences implementing AI.

"There is no silver bullet," John Woods, provost and chief academic officer at the University of Phoenix, told attendees of the education and technology conference on Monday. "You've got to lead with strategy."

Strategy-first implementation

Lisa Marsh Ryerson, president and CEO of Southern New Hampshire, emphasized that college officials should be guiding AI, not the other way around. That means that the primarily online nonprofit institution is not simply buying a bunch of product licenses for its over 189,000 students.

"AI is a help, of course, but we should not have it build our strategy," she said during Monday's panel. 

Woods agreed.

"There's 40 companies out here who will hand you a business card and say they can put a tutor in every class," he said. "It's not quite that easy."

Before hitting the marketplace, college officials should have a clear vision of what they want from an AI tool, both in terms of institutional outcomes and what its adoption would mean for learners and instructors, according to Susan Winslow, CEO of Macmillan Learning, an ed tech company that offers AI tools.

From there, they can consider acquiring tools — and beefing up their engineering staff — to support those goals, Ryerson said.

Selecting an AI tool is "not a one-and-done" conversation, Winslow said, especially in light of the potential cybersecurity risks AI poses. Administrators should regularly check in on their AI usage to ensure it aligns with their strategic priorities.

Classroom support and tutoring tools

One of the most popularly advertised applications for AI is increased one-on-one instruction. Dozens of AI-focused companies have begun offering tutor chatbots in recent years, with sector heavyweights like Google and Khan Academy launching their own versions as well.

But simply telling students they can now access an instructional chatbot will likely fail to reach those who could benefit the most from the help, according to Woods.

More often than not, the students who use tutoring tools already earn strong grades — those looking "to keep their As or move their B to an A" — he told conference attendees.

Instead of introducing broadly available help, Woods recommended implementing AI tools that complement a course's curriculum.

At the University of Phoenix, a for-profit institution with over 82,000 students, students studying to be mental health counselors use AI chatbots to practice client conversations. Faculty then review those transcripts and provide feedback, he said. Education majors similarly use AI to practice conversations that may arise in parent-teacher conferences.

In both cases, human oversight is central to the process, and students are "not just using AI for AI's sake," Woods said.

The prospect of personalized education — such as offering a law student and a nursing student custom instruction in the same history class — is increasingly becoming attainable from a technology perspective, according to Woods. 

But he added that most colleges are not set up to execute on that vision and would need a strong grasp on what each student needs and when they need it, he said.

Younger job seekers are more concerned than their older counterparts about AI’s effect on their skill sets and education, according to an Indeed report published April 21. The job site commissioned a Harris Poll survey of 772 U.S. adult workers and job seekers with an associate’s degree or higher for the report.

Businesses show no signs of slowing AI adoption, and already almost half of Gen Z job seekers say the technology's adoption has made their college education irrelevant, compared with about one-third of Millennials. About 1 in 5 Gen Z and baby boomer survey takers say the same of their college education.

The ongoing push to deemphasize college degree requirements in job postings has led half of Gen Z job seekers to view their degrees as a waste of time and money, according to the report.  

The deluge of AI adoption in the workplace has influenced in-demand roles, salary trends and even the C-suite’s purview. The effect is far from over, with younger generations grappling with the potential for additional disruptions ahead.

Every job currently posted on Indeed's job board will likely experience some level of exposure to generative AI and the changes it represents, according to Indeed Senior Talent Strategy Advisor Linsey Fagan.

"For any organization to succeed with AI, every single employee needs to have a basic understanding of AI and how their company uses it," Fagan said in an email to CIO Dive. "Leaders play a crucial role in this shift by assessing their teams, listening to individual needs, and supporting their development."

One way CIOs have responded to the wave of change is by creating upskilling programs. Technology vendors, eager to remove the obstacles toward enterprise adoption, have also deployed training programs open to the general public.

Online learning and training platform O'Reilly found skyrocketing demand for AI-specific training last year, according to a January report. The number of professionals seeking training on AI principles more than quadrupled last year, the firm said.

"To truly unlock the potential of AI, organizations must invest in their people, offering training, hands-on experiences and opportunities to explore new tools in a supportive environment," Fagan said. "Organizations need employees to be motivated to try these tools and want to apply them in their day to day."

Ever since the public release of ChatGPT in November 2022, artificial intelligence has dominated conversations related to higher education and the future of work in the U.S. Now, some colleges are investing significantly in AI-related programs, from specific degrees to integrating AI literacy into other disciplines.   

They are doing so for several reasons. 

Those include responding to predictions that the American workforce will rely on AI much more in the future. For students who would like to work with the development and science of AI, that can mean jobs — some of which are fairly high-paying. For students in other disciplines, that could mean they need to demonstrate AI-related knowledge or competency to land jobs. 

“I don't know how higher ed could not embrace AI literacy as a learning outcome,” said C. Edward Watson, vice president for digital innovation at the American Association of Colleges and Universities. “We can't not lean in to provide students with the skills that we know they're now expected to have when they graduate.”

But institutions may also be responding to student interest in AI. As many institutions deal with enrollment declines, that can be a powerful motivator. 

“There are only a finite number of students and they're going to decide where they're going to go based upon which institutions will best prepare them for life beyond college,” Watson said. 

How are colleges approaching AI?

So far, around 100 universities have launched some sort of credential in AI, said Luke Koslosky, a research analyst with the Center for Security and Emerging Technology. 

And interest appears to be growing around the topic. Conferrals of AI-related undergraduate and master’s degrees — which include computer science and cybersecurity — have increased about 120% since 2011, according to a 2023 report from CSET. In 2021, colleges conferred over 140,000 bachelor’s degrees and roughly 70,000 master’s degrees releated to AI. 

AI credentials typically live within computer science or engineering disciplines. The University of Pennsylvania, for example, offers both bachelor’s and master’s engineering degrees in AI. And some institutions offer AI-related minors. 

Other colleges are taking a different approach — integrating some AI science or AI literacy into other disciplines. 

That’s how the University of Florida is approaching its AI initiative, which aims to integrate AI into every undergraduate major and graduate program. So far, the university has incorporated AI education into a handful of undergraduate programs, including majors in information systems and data science and a minor in bioinformatics. It’s also added AI education to several graduate programs. 

Additionally, University of Florida offers a general AI certificate and discipline-specific AI certificates in hospitality, geography and public health. 

The AAC&U is also embarking on a project to help departments integrate AI into their programs. The higher education association is working with 124 institutions through its Institute on AI, Pedagogy and the Curriculum, launched in September. 

Departments submitted proposals about what they would like to achieve as far as integrating AI literacy into their curricula to receive assistance through virtual events and mentorship. Teams had to involve at least five people on a campus, preferably with one senior leader. 

“It was becoming arguable that maybe in addition to being able to write well and speak well and to be a critical thinker that additional, essential learning that should occur in college may have to do with developing AI literacies,” said Watson, the director of the institute.

The process can be difficult because faculty need to be competent in AI. Then, teams need to decide in which classes they’re going to teach AI literacy, what assignments they are going to give, and whether to focus on AI generally or in a more discipline-specific way. 

Choosing to invest

Despite the growing interest in the emerging technology, investing in AI-related programming is often difficult. For one, depending on the level and focus, it can be expensive. AI curricula can require colleges to hire qualified faculty and staff and pay for significant computing power. 

That’s why many of the institutions that are investing in AI, such as Carnegie Mellon University and Massachusetts Institute of Technology, are well-resourced with large endowments, Koslosky said. 

Other institutions are pursuing partnerships with industry to make their goals possible. University of Florida, for instance, has a partnership with chipmaker Nvidia, which includes a $50 million gift from the company and one of its cofounders. Arizona State University is partnering with OpenAI, to provide enterprise subscriptions to ChatGPT for approved faculty and staff.

Stony Brook University, part of the State University of New York system, recently expanded an AI institute into a universitywide initiative, which will focus on research and applications in healthcare, infrastructure, education and finance. The university is investing about $15 million, which includes support from Empire AI, a college research consortium focused on AI and heavily funded by the state.

Stony Brook Provost Carl Lejuez said that “$15 million is nothing and a lot of money all at the same time.”

“We’re seeing companies struggle with this immensely because they’re having to make decisions about — are they going to fall behind so far that they’re not going to be able to compete?” Lejuez added. “But in the meantime they’re spending millions and in some cases billions in cases where they’re not generating real revenue yet.”

Whether an investment in AI programming is right for an institution will likely depend on its circumstances, including its resources, faculty, mission and connections to industry.

“If you're a major research university, you're going to be really far behind if you are not investing in this,” Lejuez said. “For research universities, it's an absolute necessity.”

For other institutions though, a big AI investment might be riskier.

“If your school doesn't have a huge computer science department and doesn't have a lot of industry connections to companies using AI or building AI, then you shouldn't drop everything to stand up a brand new AI program right now,” Koslosky said.

Additionally, investments in AI are so far untested. Although many business leaders believe AI is going to change the American workforce, that transformation hasn’t fully come to pass. Although current research predicts the number of AI jobs trending upwards, what those jobs look like might change. That means that trying to integrate AI into other disciplines is the financially safer approach, Koslosky said.

“Schools are struggling with lots of things and balancing competing priorities,” he said. “AI is important and will continue to be, but it's not the only important thing.”