Why human touch is still essential in an AI world

Like most of you I’ve been following the accelerated interest in and discussion about AI. This weekend I was fortunate enough to sit in on a few sessions at the Sydney Writers Festival featuring the eminent AI researcher Toby Walsh.

A few thoughts.

We need to recognise that AI is not new. It has been around for a very long time and it is already very present in our daily lives – the algorithms in google maps or speech recognition in Siri are obvious examples. At StudyLink we’ve been using AI in our products since we developed our first course search algorithm on our CD ROM in 1996 and then moved it online to the studylink.com web directory in 1998.

What is new is generative AI and the ability for the automatic creation of content and response and the rapid improvement of this since the launch of chatGPT in late 2022.

The A in AI is ‘artificial’ and while exciting there are limitations. Netflix uses AI for its recommendation engine. And yet, for my viewing selections I also rely, and maybe more so, on recommendations from friends. The Netflix algorithm helps but it’s not my main source.

There are issues with the removal of humans from the process and the complete reliance on machines. This has life threatening consequences in the use of AI in weapons but is also relevant for international recruitment. The removal of this direct human relationship between a university and their agent is one of the main concerns with aggregators. Removing the visibility of which agents are recruiting for institutions raises questions about the quality of counselling and guidance that has been given and whether it complies with the various codes, regulations and good practice. It remains to be seen whether the sophisticated technology supporting aggregator processes is sufficient to ally these concerns.

We need to also be aware of the issues of removing humans from the counselling and admissions process. AI plays a role in helping direct and choose but you cannot replicate the knowledge, empathy and emotional considerations that a person adds. There is acknowledged biases in AI that can only be detected with human intervention. Recent class actions in the US against AI enabled CV filtering tools have only highlighted the concerns our industry should have about the wholesale adoption of AI in counselling and admissions. Feedback from our admissions users is they want the technology to deliver efficiencies. To help them prioritise and automate where possible so they can focus on what’s important and don’t need to work long hours and weekends. But human touch is still key, no matter how quick or small, to make sure the machines aren’t excluding someone they shouldn’t. Internally we refer to AI as ‘Admissions Insights’ and our recent Priority Assessment tool is an example of one of the many available in StudyLink Connect.

The opportunities to use AI for improving decisions and saving time is exciting. But we also need to be careful and not be distracted by anything promoting the use of chatGPT or AI as automatically good or ‘must have’. I’m easily distracted by the latest shiny ‘tech’ offering but luckily our team is very patient. They take on board my suggestions about the latest developments and then listen to our users about what is needed so they can develop tools that make sense and have maximum impact. They’ve been recognised for this recently by being shortlisted for a PIEoneer Award after developing a simple tool to easily verify English language results at source that saves thousands and thousands of hours.

Exciting times to come…

Published by Jason Howard

StudyLink Connect, Founder and CEO

In 1992, while still at university, Jason founded StudyLink Connect. Initially a CD ROM multimedia guide, StudyLink Connect is now a leading international recruitment channel and admissions management solution. During that time, Jason has helped many providers understand new technology and how it can be used effectively in their recruitment plans.