Researchers from the University of Colorado Boulder (CU Boulder) and the University of Pennsylvania have unveiled a remarkable AI tool that can extract crucial personal traits from college application essays, Phys.org shares this report.
This groundbreaking technology could revolutionize the admissions process, offering a more holistic view of applicants beyond grades and test scores - or is this relying on the technology too much?
CU Boulder and University of Pennsylvania unveil AI tool analyzing college application essays, revolutionizing admissions process with a holistic approach, addressing algorithmic bias.
Unveiling the AI Marvel
College admissions are highly competitive, and thousands of applicants vie for limited spots, institutions have long grappled with the challenge of evaluating the personal qualities of applicants. While grades and standardized test scores provide essential metrics, they do not capture the full spectrum of a student's potential.
The team behind this AI tool has designed it to analyze college application essays and identify qualities such as leadership and perseverance.
This approach leverages artificial intelligence to provide a more comprehensive understanding of applicants. The study emphasizes that AI can enhance the holistic admissions process, which considers applicants as individuals, not just numbers on a page.
The study, led by researchers from CU Boulder and the University of Pennsylvania, was published in the journal Science Advances in October.
AI with a Conscience?
One of the most significant aspects of this study is the team's commitment to avoiding algorithmic bias-a common issue plaguing AI systems, particularly in the context of facial recognition software.
The researchers have worked to ensure that the AI tool does not exhibit 'algorithmic bias,' a preference for applicants based on their racial or gender backgrounds. This commitment to fairness and equity is a promising sign for AI in admissions.
It is important to note that some studies have highlighted the tendencies of AI to be biased against a certain group of people.
Incorporating Transparency and Expertise
No, the AI tool is not intended to replace human admissions officers. According to study co-author Sidney D'Mello, the tool should serve as a supplement to human decision-making, not a replacement.
It can assist admissions officers in identifying promising students who may have been overlooked in the deluge of applications. The researchers advocate for transparency in AI decision-making, ensuring that the AI provides explanations and communicates its confidence level in its assessments, allowing people to make informed decisions.
Read Also: Google Unveils AI-Powered Ad Creation Tools for Agencies, Businesses
Data-Driven Insights
The research team, in a quest to make the admissions process more reliable, tapped into a substantial dataset of over 300,000 college applications submitted in the U.S. in 2008 and 2009. Each application included a 150-word essay where applicants described their extracurricular activities or work experiences.
Real admissions officers and undergraduate students were engaged to evaluate these essays, identifying evidence of seven personal qualities that colleges often seek in incoming freshmen. These qualities included intrinsic motivation and prosocial purpose, which measures a willingness to help others.
AI's Incredible Role
The research team then harnessed the power of AI, specifically large language models, to learn from these human ratings. They successfully trained the AI to identify evidence of personal qualities beyond mere keyword spotting. When these AI platforms assessed new essays, their results closely mirrored the judgments of human readers.
Notably, the AI demonstrated an equitable distribution of beneficial personal qualities among applicants from various demographic backgrounds, with female writers slightly more likely to exhibit prosocial purpose.
Perhaps the most surprising revelation was the profound impact of personal traits on college graduation rates. Students whose essays exhibited evidence of leadership were found to be more likely to graduate within six years, even after considering factors like test scores and demographics.
The study emphasizes the significance of the language embedded in these essays, revealing that it can serve as a reliable predictor of success in college.
Stay posted here at Tech Times.
Related Article: AI Text Detector Achieves Nearly 100% Accuracy in Spotting Human-Written from AI-Generated Scientific Papers
