Artificial Intelligence in Student Admissions - What AI can and can't do...

I've been having a lot of discussions with universities lately about the concept of using Artificial Intelligence (AI) in the University Admissions process. It is a attractive concept for many universities because their admissions processes are often the poster child of inefficiency. Typically University Admissions process large volumes of applications which require many individual documents to be sighted and verified. Often many different processes are involved that have many steps, many options, multiple sub-processes and special considerations that all need to be addressed and can vary depending on the degree program and course structure. Most of these transactions have to be processed over relatively short periods of time two to three times per year at the beginning of each semester when the universities do their intakes which puts even more pressure on university admissions teams. This makes admissions a complex and challenging area for university operations.

These challenges are increasing as the competition between universities intensifies for the top students. It is well known that the longer it takes for a university to issue it's offer to a student the lower the probability that the student will accept that offer. Most students actively shop around to many different universities (likely in a number of different countries) to get the best offer. They typically adopt a "bird in the hand" philosophy when choosing a school to go to, accepting the first suitable offer they get.

So universities are keen to improve their admissions processes, increasing speed of processing and reducing time to offer with a view to getting more of the top quality students in their doors.

Why now with Artificial Intelligence?

Artificial Intelligence has been around for a number of years. But it has increased in popularity in the last few years. So much so that Gartner (2017) now boldly stating that "By 2020, 85% of CIOs will be piloting AI programs". But why the recent popularity? Well there are a few reasons....

First, we have reached a tipping point in the amount of available data. AI is heavily reliant on the volume, velocity and variety of available data. As each of these factors increases the better AI gets at predicting an outcome and making a recommendation based on that outcome. With the advent of the Internet and now the Internet of Things the amount available data is growing exponentially.

Second, with the advent of cloud computing the availability of large amounts of low cost computing make AI more affordable. AI traditionally requires a large amount of computing power over a short period of time to train the mathematical models used in AI. Now with cloud computing offering the ability to purchase extremely large amounts of computing power, cost effectively for a short periods of time, AI becomes more practical.

The third and final reason for AI's recent surge into the limelight is the advent of better math and better tools to deliver the math. Recently improvements in the mathematical models mean that AI is getting more accurate in it's approximation of reality. Answers from bots are more human like and AI models more robust at handling a variety of inputs. Also the tools available to data scientist's and programmers to build AI's have evolved to the point where building and delivering an AI can be done in days. I was having a play last week using the AI tools from a well known software vendor and was able to get an AI model built and deployed as a web service within a few hours. Now it was a relatively simple AI model that I built, but my point is that these tools and techniques now make it practical for most IT departments to build and maintain AI tools themselves. If I can do it they can do it too! :)

So AI is definitely within the reach of most university IT departments at the moment.

Narrow AI vs General AI?

One has to be careful when discussing AI to not paint it as a panacea. In AI there are widely held to be two types of AI. There first is "General AI". This is where we build a machine that thinks and acts like a human in every way (e.g. the android call Data from Star Trek).

The second is called "Narrow AI". This is where AI is used for specific processes or purposes like recommending music to you on Spotify, identifying what shows you'll be interested in for Netflix or figuring out how to respond when you ask your Google Home what time it is.

The reality of it is Narrow AI is here and now being used by many organisations quite effectively. General AI is probably 20 years off.

My point here is that when deploying your AI technologies you are well advised to deploy them for specific processes keeping the application as narrow as possible at least initially. As you train and deploy your AIs for those specific processes you can then broaden their application and slowly increase their range of applicability.

Use Cases of AI in University Admissions

So we can say that many University Admissions groups are in need of help, that AI is a practical technology to deploy now and that if we keep its application to specific uses cases our chances of success go up immensely. So what are some of the use cases where AI could and has been used effectively University Admissions. Well the ones I have been talking to universities a lot about recently include...

• Identifying fraud in admissions applications;

• Predicting application acceptances by students and thereby identify ways to improve acceptance rates and to improve capacity planning;

• Document recognition, validation and approval;

• Scanning for and identifying policy violations;

• Augment your admissions processes by making recommendations regarding specific applications and flagging applications that need follow up for various reasons;

• Chat bots to take inbound inquiries, answer questions related to ATAR scores and helping to build course schedules; and

• Many many many others.

Conclusions Regarding AI in the University Admissions process

AI is quickly expanding its use in university admissions across the board. It will be applied to University Admissions for a variety of use cases in an increasing and every expanding way in the near future. This is not just me saying this but some very well know senior university leaders as well. Check out a recent speech done by a University president out of the USA: https://news.elearninginside.com/university-president-predicts-ai-will-takeover-the-university-admissions-process/ . These improvements in University Admissions efficiency will no doubt mean improvements in service levels for the students which is a win/ win for everyone involved!