One day, Kinjal Chaudhari, a 2026 University alum in Engineering, and Mansi Vaghela, a 2026 alum at the University of Illinois Chicago, had an idea. 

They, along with some colleagues, used artificial intelligence to completely generate a research paper. 

Chaudhari said the large language model produced a paper with equations and graphs that looked like a novel. It underwent publication after two rounds of peer review, according to Chaudhari. 

Turns out, the pair of students, along with colleagues across multiple universities, had a bet on whether AI could write a research paper and get published — and it would have, if the group hadn’t pulled it back. 

Chaudhari hit the alarm button. 

“We’ve been told throughout our entire education system that our academic literature is our ground truth,” Chaudhari said. “At the end of the day, if a paper has fancy enough vocabulary and complex equations, it’s going to make it through, whether it’s right or not.”

Chaudhari said that experience inspired her and Vaghela to co-found 20/20, a semantic claim verification software that helps reviewers evaluate manuscripts for accuracy, especially for inaccuracies caused by AI use.

Chaudhari and Vaghela say they’re starting with academic peer review and research papers, but hope that one day, their software can help lawyers and consultants make high-stakes decisions. In a world where AI makes it “easy to stop looking,” they hope to give reviewers the context they need to determine what is credible.

Chaudhari co-founded the company as a student at the University, which allowed her to partner with iVenture Accelerator, a University program to help startups with funding and mentorship. When Chaudhari started, she said the project was an attempt to solve a genuinely interesting problem — there was no intention of making a business, which is where iVenture came in. 

“Having a structured, safe program to really make all the mistakes and figure out how it works and have advisers to help you was definitely beneficial,” Chaudhari said. 

While Chaudhari, the chief technology officer, leads the model development, Vaghela is the entrepreneurial lead for 20/20.  Vaghela emphasized that research published fraudulently has implications beyond the immediate subject. 

According to Vaghela, a decent proportion of research that scientists, doctors and engineers rely on contains traces of potentially fraudulent work.

A study published in 2025 in Nature Human Behavior found that up to 22% of computer science papers were used in LLMs. So how does 20/20 plan on verifying the veracity of papers when AI seems so ubiquitous? Funnily enough, and perhaps to one’s relief, not through AI.

“Fundamentally, AI cannot police AI,” Chaudhari said.

Instead, Chaudhari said 20/20 uses statistics to compare submitted papers with existing literature to flag similarities, observes how the two resemble each other and helps reviewers verify that the claims made are supported.

For Vaghela, the next step was to put the software to the test in the field. The two generated fraudulent papers and asked researchers to peer review them using the software.

“We found that all of the reviewers said that our tool helped them catch fraudulent claims that they would not have seen otherwise,” Vaghela said. “That was a proof of concept and was motivating for us to hear. There’s actually demand, and it’s actually useful.”

While Chaudhari handles the code, Vaghela manages outreach. Vaghela said the worst thing a founder can do is “build in silence,” so she networks and works on customer discovery. These conversations with customers are what Vaghela labeled as her proudest moment: when they understand how the product works and think they can integrate it into their existing workflow.

Despite the success the pair has seen, the uncertainty of running an early stage company as a recent graduate is nerve-wracking, according to Chaudhari.

“You see all of your friends starting their full-time jobs at corporate companies, moving away to different cities, and then you’re investing in yourself, right?” Chaudhari said. “So that is a big sacrifice to make. But at the end of the day, you can’t just give this a little bit of effort. If you’re going to go full in, you have to go full in.”

Nevertheless, Vaghela pointed out a benefit to taking the risk early. She said starting a company out of college is easier than giving up a full-time job to begin. 

“There’s definitely a lot of sleepless nights, and you have to put all of your time and energy into something like this,” Vaghela said. “But if it’s something that they’re passionate about and that they think needs solving, and they think they’re the ones to do it, then I think it’s a great idea to go all in.”

[email protected]