California Community Colleges are all in. From their (now) annual Futures Summit on September 10 to their work with AI in fighting admissions fraud, the 116 colleges of the California Community College system plan to be front and center in the use of AI Technologies.
One factor that has really moved the colleges along is the very high incidence of admissions fraud, which according to the state chancellor’s office, is dealing with some very troubling numbers. In the past year alone, more than 460,000 requests to the state’s online application system are fraudulent requests from bad actors. That adds up to a $7 million loss this year alone, and enough is enough.
It seems that bad actors have found an opening, and they are employing some very sophisticated technology to steal their bounty. But the community colleges are not taking this theft lying down. Luckily, they have Kiran Kodithala and his company, N2N in their corner.
Before Kiran and company, the colleges were trying to do everything manually, and the manpower required to stop the fraudsters was overwhelming. But N2N uses AI to fight fire with fire. And it’s working.
According to Sai Guptha Grandhi, a data scientist at N2N, “Machine learning (ML) is particularly effective in detecting fraud because of its ability to analyze vast amounts of data quickly and identify hidden patterns that would likely go unnoticed through manual review. Traditional methods often rely on simple rule-based checks (like flagging repeated IP addresses or age discrepancies), which are limited to catching only the most apparent signs of fraud. ML models, however, excel at finding nuanced connections—such as linking shared phone numbers, addresses, or email formats across different applications. This triangulation method reveals networks of fraudulent activities by uncovering correlations between seemingly unrelated applications, thereby offering a more comprehensive defense against sophisticated fraud tactics.”
N2N’s system outperforms the traditional methods the colleges were using due to its speed, accuracy, and adaptability. Unlike manual checks or rule-based systems, the ML model continuously learns from new data, evolving to detect emerging fraud strategies. The system employs a triangulation approach, cross-referencing multiple data points from applications to uncover hidden links. For instance, N2N’s model identified a 2x increase in detecting fraudulent actors, as noted by their partners, with a 92.3% effectiveness rate upon redeployment at the West Valley Mission (WVM) Community College District. Additionally, they achieved similar accuracy at Santiago Community College with a 24 hour turnaround, highlighting the system’s efficiency.
“Our model is designed to adapt and improve over time through continuous learning,” said Grandhi. “Recently, one of our beta partners identified false negatives—cases where fraudulent applications were mistakenly approved. By incorporating these cases into our model’s training dataset, we enhance its ability to detect similar fraud patterns in the future. We also collaborate closely with our partners, refining the model by adding new features (like phone number histories and geolocation data) and removing outdated predictors. This iterative process allows the system to scale efficiently, transitioning from working with a few colleges to serving multiple institutions statewide, providing stronger fraud detection with each new deployment.”
According to Kiran Kodithala, CEO of N2N, “it took less than 3 years for AI to take off. This is the first time that anyone from five to 80 has been able to take a platform and communicate with it without needing any training whatsoever. And that’s what makes AI truly powerful.
“As we began to use AI, we noticed there was a big disconnect between how easy AI is and how hard college systems are in a lot of ways. And not just the classroom instruction. I’m talking about the systems that students have to negotiate on a daily basis. The student information system that is the authoritative source for student data, faculty data and other data learning management systems that act as the authority source for different learning content, and other administrative systems, like the parking, housing, and others, each of them goes through myriads of logins, security profiles, and technologies. So, one of the big issues we were trying to figure out is, how do we give that type of flexibility that chat GPT has, or Gemini has, where people can just search and get the answers, and that’s what led us to a generative AI product by N2N Services. I personally started my journey as a student information system implementer between 1998 and 2010, and since 2010 we’ve been focused exclusively on connecting different systems student information systems, learning management systems, parking systems, structured and unstructured data.
“And as a result, we built a product called Illuminate, which is an integration platform as a service that allows these disparate systems to be accessible, using the power of APIs or application programming interfaces.
“What we’ve done to make LightLeap AI work is we converted each of these into AI agents so that students can ask questions to the bot, and it will answer them based on their profile. The big difference on how we built our AI product compared to other AI products is, there’s this idea by some companies or product vendors that for you to make AI work, you have to give all your data to a vendor, and the vendor will build an AI product for you. We believe that the only way AI can be successful is to keep the data in the authority sources and plug it in for specific point releases using security. So, when a student says, ‘Where is my course Information,’ it’s not pulling it from my database, it’s pulling from your own student information system.
“A lot of this comes from the hard work from my team. But really, the only reason why we’ve been able to grow this and implement this at this pace is because of our fantastic collaboration with our lead partners. Ivy Tech Community College has led the way in implementing a cool transcript ingestion system where a student literally uploads a transcript and it processes the transcript and shows the result in a text format and pushes it to banner. Foothill De Anza has been leading the way in building a fraud detection model where we are identifying frauds almost in real time and zapping them from registering for courses and applying for financial aid. And West Valley Mission Community College District has been leading the way in building a chatbot that gives students the experience that they can be most effective with, and Rancho Santiago Community College has recently joined us to guide the way on how we can use AI for stopping financial aid fraud across the board, and we also have a great collaboration with the California Virtual Campus, where they want to build a plugin to help a faculty member identify where the areas of opportunities are to improve their lessons and make it more RSI compliant.
“At the onset of Covid, so much classroom learning shifted from in person to online. That was one factor; the other thing that was going on was the Equifax and T-Mobile breach, in which millions of people’s personal information were getting stolen and hosted on venues like the dark web which gave bad actors like fraudsters the ability to steal that information and apply to campuses and then register for classes, preventing actual students from being able to enroll in those classes and then going on to apply for financial aid, never showing up for those classes and then skedaddling with the money.”
According to Jory Hadsell, Vice Chancelor and Chief Technology Officer at Foothill De Anza Community College District, we were really pleased to work with the team at N2N to do the initial development of the AI fraud detection model. Like many institutions, especially open access institutions in higher education, we were devoting a significant number of resources and staff time specifically to reviewing applications for fraud. It’s been really incredible to have the model now deployed. We ran a lot of historical data through it, did some fine tuning, worked a lot with our financial aid and admissions teams, and I can’t thank them enough for the time they spent reviewing different reports and helping us to tune the model. It has significantly reduced the amount of time that they have to spend and has been incredibly accurate. Of course, no system is perfect. We thought we had a fairly advanced model in place. We were one of the districts that did have some automation built around this, and found double the amount that we thought we had. So, we’re really happy to see this scaling. And every institution that adopts, is making the product even better, and we all benefit from that.”
Admissions fraud is a significant problem nationwide. But the California Community Colleges are fighting back. AI warriors like Kiran Kodithala and his team at N2N Systems are helping them win back territory, saving millions of dollars and freeing up seats for actual students. Admissions fraud may already be a $1 billion problem, and there will be a lot more losses before good guy warriors can fully stem the tide. But they are on it. And with the help of rockstar administrators like Jory Hadsell and others, the fraudsters had best watch their A**es. Because Kiran and the California Community Colleges are now kicking some serious AI.
About the author
Kiran Kodithala is a technology leader and a seasoned expert in the field of educational technology. As the founder and CEO of N2N Services Inc. since 2010, Kiran has been at the forefront of driving innovation in higher education through technology.
With a career spanning over 25 years dedicated to the sector, he has established himself as a pivotal figure in transforming how educational institutions leverage technology for better outcomes.
Under Kiran’s leadership, N2N Services Inc. has grown to become a leading integrator providing cutting edge solutions to the challenges faced by educational institutions. His passion for integrating technology into education has led to the development of transformative products like Illuminate and LightLeap.ai. These innovative platforms embody his commitment to harnessing the power of AI and secure data integration to revolutionize educational institutions globally.
