Lily Pratt, PHAA Policy Officer
It has been another incredible turn-out for the 2025 PHAA policy and advocacy intern program with almost 300 applications vying for several spots at PHAA, Cancer Council, Dementia Australia, and Lung Foundation Australia.
Typically, the Policy and Advocacy team each January reviews every application (a resume and answers to three prompts). Although a large task, it is usually enjoyable reading through the mass of energised and passionate applications from students and graduates. Their determination to change the world is infectious – the perfect pep talk for the start of the year.
However, this year the Team has noticed something a little… “artificial”.
It’s hard to exactly quantify the number of applicants who utilised AI software for their applications, as applications for intern positions do tend to have a certain flow, tenor and vocabulary.
Our best guess is that as few as one third and as much as one half of the applications we reviewed had almost the EXACT same phrases, order of information and/or non-specific generality inherent in ChatGPT responses. There were also a few unfortunate cases where applicants forgot to remove the ChatGPT marker from their answer, e.g., “ChatGPT Here’s a response tailored to your interests:”.
Reading through the generated responses was a bit, well, sad. Previously, applications were rich with real examples of personal experiences in public health, or with passion as they describe an injustice they’ve studied or seen, or even with vulnerabilities they feel about the certainty of their career direction. It is a painstaking task to refine the applicants as most are incredibly deserving.
Understandably, ChatGPT is a software, so it cannot express itself like a human would. There are instances where using AI is arguably not a problem, but a benefit. Mundane writing tasks, like bulk emailing, has been cited as an example where AI has helped in improving efficiency.
However, public health is all about complex human issues, experiences and behaviours; how we experience the world, how the natural, commercial, and built environment impacts us, understanding and questioning the structures and systems that humans have built to organise ourselves. When we ask, “what interests you in a public health internship?” (prompt one), it is a very personal question that only a human can answer properly and uniquely.
No matter how much one “tweaks” an AI generated response to such a question, it’s just not quite the same as what the good old fashioned human mind can do.
Given the volume of AI responses, we are left to consider how to approach this growing challenge. Do we decline each application identified or suspected as using AI? Or do we accept this “norm” that our academic members have been navigating for a few years.
On one hand, to reject all potential AI applications could mean accidentally penalising applicants who did write it themselves but just have a particular writing style. Also, as my colleague Malcolm Baalman thoughtfully said, given the difficulty of finding a job right out of university, this might have been the 10th position people had applied for that day/week. Can we really judge so harshly when people might want to streamline this one process amid living a busy life?
On the other hand, if we accept the use of AI generated applications as the “norm”, we can’t accurately assess someone’s ability to express an idea, nor can we assess someone’s true interest in public health policy. Also, we’d have many applications that read almost identically, so cannot be judged separately anyway. There’s also the unavoidable fact that we look for depth and expression in the responses to the three prompts, which a typical ChatGPT response generally does poorly.
So, what did we do?
Without having to purposefully cast out any AI generated applications, the highly personalised applications were just more likely to progress through our refining process. That’s because they were better at communicating to us who the applicant was, and why they felt the internship could help them. Also, they were more likely to show critical thinking and they made us want to learn more about the person behind the application.
I think that in today’s world of misinformation, and doctored images and videos, any technology that encourages us to think less is not a positive. Writing about oneself can be difficult, but it is a valuable way to practice reflexive and reflective thinking – which we should be encouraging despite the uncomfortable nature of introspection.
As to whether people should use AI to apply for jobs and internships? Well, let’s ask AI itself:
“Using ChatGPT to help with your job application can be a great idea! I can assist you in structuring your application, improving your language, and suggesting ways to highlight your skills and experiences.
If you’d like, I can help you draft a cover letter or review your resume—just let me know!”
Never mind.
My reactions, and those of my colleagues, to this same question are consistent; we don’t recommend using AI to answer any question seeking to determine what you really think, and who you are. Nothing beats your voice and your expression. That’s what potential employers want to hear in your application.
If you need help structuring your application, there are some excellent resources online that provide examples, templates and tips.
- GradConnection
- Tips for how to write an application for an internship
- How to Write a Cover Letter for the Australian Market in 2025
- Harvard Business Review: How to Write a Cover Letter That Sounds Like You (and Gets Noticed)
Leave a Reply