Key points:
- AI offers both unprecedented efficiency and the risk of perpetuating biases in admissions
- How universities ethically leverage AI
- College counseling in an AI world: A path to fair access
- For more news on AI in admissions equity, visit eCN’s AI in Education hub
As colleges and universities navigate the shifting landscape of higher education, the appeal of employing artificial intelligence in admissions processes has grown significantly.
These institutions are grappling with multiple pressing challenges:
- Maintaining diverse enrollments in a post-affirmative action landscape
- The unavoidable demographic cliff
- Ensuring equitable access
- Managing the escalating administrative burdens due to increased applicant volumes
This complex environment demands innovative solutions to streamline operations without compromising the commitment to institutional goals such as diversity, equity, and inclusion (DEI). When thoughtfully implemented, AI offers a promising avenue to address these multifaceted concerns, enabling more efficient review processes and potentially more nuanced decision-making that could lead to more balanced and diverse classes.
Harnessing AI in college admissions: Navigating efficiency and equity
Integrating AI into higher education can make the whole process more systematic and streamlined. Already, the technology is being tasked with evaluating application materials and screening everything from transcripts and test scores to letters of recommendation and essays. With the number of applicants now on the rise, increasing by 32 percent between 2020 and 2023, automation can shoulder much of the administrative load.
The problem is that the DNA of the data being fed to AI is biased. Data comes from people, and people are inherently biased. Even an unconscious bias can cause AI models to drift, unintentionally introducing prejudices toward certain groups or perpetuating biases and thus leading to greater disparities among gender, race, and ethnicity. So, how can we use AI in college admissions most effectively to ensure the process is equitable?
First, we must change our mindset about technology. There are two sides to the current spectrum: AI is either viewed as the enemy or held to a much higher standard than humans. People want to view it as an infallible entity, which isn’t the case. Like any technology, AI is a tool that makes our jobs easier. And we shouldn’t hold it to a higher standard than we hold humans. Think of AI as a team member–when it gives an idea or solution, check that it makes sense. Does everyone around the table agree with what AI is saying? If not, evaluate why that’s the case and hold AI accountable for its decisions.
Principles of sound practice tell us that if we put 20 people in a room with diverse backgrounds, the chances of an unbiased outcome grow exponentially. Though each person comes to the table with their own biases, the different perspectives put everyone in check. That’s how AI in college admissions should work–as a part of a diverse team where one member isn’t making all the decisions. AI offers a different perspective, and we should consider that during the admissions process.
The pitfalls and possibilities of AI in equitable admissions strategies
Along with this change in mindset, we must come to a better understanding of where AI and its application could go wrong. But that’s the great thing about AI and machine learning–the technology allows us to diagnose the issue before developing a solution. We can tell AI to identify where a bias might exist. Perhaps it’s in the way we collect data. Maybe it comes down to how a model was built or the data used. Or maybe the team isn’t using the insights derived from the data appropriately and is introducing biases after the fact.
Take something like an engineering program, for example. At many colleges and universities, it’s hard to deny that there’s a need for more women in these programs. Why is that? Is it because of the data? Is it because of our preconceived notions that women aren’t interested in the field? Historically, white men pursue this type of degree. But if we were to take the data, build predictive models, and then flip the gender, how would the probability change? Simply flipping the gender can help us detect bias while also providing insights into how to get more women into the program.
With resources in short supply in admissions offices, these kinds of insights and prescriptions allow us to focus on the strategies that raise the probability of a given individual enrolling. This foresight also reduces the cost of making a prediction. AI handles all the hard work of data processing, allowing humans to do what we do best–make creative decisions. The analytics help us understand certain applicants and what it might take to get them interested in our school. For example, AI might indicate that one individual is 20 percent likely to enroll and then offer five different things we can do to increase that likelihood. We can then pick and choose which strategies we feel will be most influential in their decision to apply and enroll. Remember, AI gives a probability of the answer, not the end-all-be-all solution.
Using AI and data analytics to better understand certain households can boost equity in college admissions. The old ways of thinking about applicants simply in terms of test scores and GPAs are gone (or should be). Building a balanced class starts with building relationships with households (e.g., parents and guardians) from diverse backgrounds with early high school and pre-high school students. AI can help create this bigger, more diverse top-funnel group for admissions teams and tell us who is most likely to enroll so we know where to focus our efforts.
AI can also simplify outreach personalization. Potential students (and their parents) want to consume information as individuals; they want to feel personally connected. At the same time, admissions teams want to understand students on a personal level so we can determine whether they match the institution’s mission. AI can help simplify this future of communication.
As the higher education landscape grapples with the complexities of DEI in the wake of legal challenges to affirmative action, the role of AI in admissions emerges as a double-edged sword–offering both unprecedented efficiency and the risk of perpetuating biases. To truly harness its potential, we must recalibrate our approach to AI, not as a panacea or a scapegoat, but as a dynamic tool in a broader strategy to cultivate a richly diverse academic community.
This journey demands rigorous scrutiny of the data that feeds AI, a commitment to reimagining admissions processes, and a concerted effort to dismantle the systemic barriers that inhibit equity. In doing so, we can build a more inclusive and equitable future in higher education, where technology serves as a bridge to opportunity.
Andy Hannah is president of Liaison International’s Othot Division, helping higher education institutions achieve enrollment and student success goals through technology. Hannah co-founded Othot in 2014, propelling the use of AI and ML for enhancing student lifecycle management, enabling prediction of student outcomes and insights generation to optimize success rates.
- The pursuit of post-crisis leadership in higher education - February 7, 2025
- When adopting AI, colleges shouldn’t override human connections - February 6, 2025
- Trump deportation threats weigh on groups offering FAFSA help - February 5, 2025