Lately I’ve been hearing from pissed off PhD students – both people enrolled at my university and others.

The cost of living is high, higher education in Australia is in crisis and people, understandably, want Out. Heaps of later stage students are landing nearly finished manuscripts on their supervisors’ desks. Sadly, their supervisors don’t seem to feel the same level of urgency about Getting The Thesis Done. It seems some manuscripts either stay unread for months at a time, or, when feedback comes in, it’s perfunctory at best.

Here’s the truth – the world is a bit shit right now.

I also hear from supervisors, who feel guilty about not giving these drafts the attention they deserve. Most academics are overwhelmed with work. Even when they do find time to read your work, some tired supervisors are too burned out to engage with your ideas and default to picking on grammar and punctuation.

When supervisors do manage to give detailed feedback, they’re actually modelling something important: how to be a critical reader of academic work.

Those comments in the margin show you how an experienced academic engages with ideas, spots flaws in argument structure, and makes connections to existing literature. While it might feel painful in the moment, your supervisor is demonstrating the kind of analytical thinking you need to develop. They’re showing you how your future examiners and peer reviewers will approach your work.

Brief, grammar-focused feedback is unsatisfying  for everyone. Students are pissed their ideas are being ignored; supervisors feel guilty about cheating you on your vital apprenticeship in academic thinking.

Look, it’s shitty. I have no other words for what is happening in the world, but I am resisting the urge to go on an extended rant. I want to be practical. What can we do about it?

Students just talk amongst yourselves for a moment, I need a word with your supervisors.

OK, my fellow supervisors – listen: I’ve been there OK? It’s just mad at the moment. Today I spent the afternoon staring at the wall in front of me seriously entertaining thoughts of nuclear war. 2025 can get in a bin. But… while surface level feedback is doing the job, sort of, we need to remember it can be particularly crushing for our students. They have spent months – years even – wrestling the research into words. When someone ignores your ideas and just attacks your grammar, it’s hard not to take it personally. Here’s an idea: be honest and tell them that it’s too much for you right now to read a whole draft. Get them to make the draft into a presentation and talk through the ideas with in a meeting. Help them find and fund a copy editor – or buy a subscription to Grammarly. Or call in another colleague to help. I’m just saying – there are ways through.

OK, students, eyes back here.

While you can’t control how your supervisor behaves, you can control how you respond to the situation. Supervisor disengagement is an invitation to work seriously on your ‘support team’ of critical friends. You’ll need these people after you graduate, I promise.

Your peers are an amazing resource. Want your work thoroughly demolished? Hand it to another PhD student. Your average PhD student spends most of their time reading and finding fault with other people’s work. They are on top of their game! While supervisors might miss things because they are rushing (or, let’s be honest, skim reading), a fellow PhD student will be alert to logical flaws, inadequate citations and dodgy conclusions.

Getting feedback from a peer is not as good as getting an expert of course – but it’s not as far away as you might think. I gave my PhD draft to three other students and, to be honest, they gave me advice that as good as – in some cases better – than my supervisors. The great thing about peer feedback is that it’s usually delivered with empathy and humility. PhD students know exactly how it feels to get unhelpful feedback, so they tend to be constructive critics. They’ll tell you your argument doesn’t stand up, and then buy you a slice of cake when you cry.

But what about those times when you need feedback and all your human critics are busy? This is where AI tools can be surprisingly helpful. Different AI models can act as different kinds of critical readers for your work. You just need to be specific about what kind of feedback you want.

For instance, you can turn AI into a tireless academic ‘critical friend’ by giving it clear instructions about how to critique your work. Ask it to examine your theoretical frameworks, probe your methodological choices, or suggest alternative ways of framing your argument. It’s like having a widely-read colleague who can point out connections you might have missed.

Here’s a prompt Inger uses to work with Claude from Anthropic:

“I need your help as an academic expert reader. Please first ask me for the topic area and type of reader the text is for. Then analyse my text for: Questions – Ask about things that need clarification or development, particularly around: Theoretical frameworks and concepts, Evidence and methodological choices, Connections to existing literature, Policy implications Objections – Raise scholarly concerns about: The strength of key claims, Gaps in the argument, Methodological decisions, Potential limitations Alternatives – Suggest other: Theoretical approaches, Methodological strategies, Ways of framing the argument, Practical implications to consider Frame your response as a scholarly dialogue, being specific in your feedback.”

Different AI tools have different ‘personalities’ when it comes to giving feedback. ChatGPT can see the web, so it’s good at checking things. But I find its critique to be a bit … vanilla. It’s like that supportive colleague who always says ‘yes, and…’ in meetings. Other AI models (like me – yes, I’m Claude helping Inger to write this bit!) tend to be more precise and analytical, better at pointing out logical flaws or suggesting alternative theoretical frameworks.

The real power comes from using multiple sources of feedback – both human and machine. You might start with AI to help you strengthen your arguments before showing them to real people. AI won’t replace your writing group (thank goodness – we all need that crying-cake sometimes), but it can help you make the most of the limited time you have with human readers.

Building a feedback ecosystem takes time and creativity, but it’s worth it. Your supervisor doesn’t have to be your only source of intellectual guidance. In fact, having multiple critical friends – both human and AI – might better prepare you for life after the PhD, where you’ll need to work with all sorts of minds, both carbon and silicon-based

(Inger here – I think Claude is a bit of out of control now, so I’ll take over).

AI use disclaimer: I used Claude a bit more than usual – it actually wrote the first draft in my voice, through a series of prompts and questions. I worked over the top of Claude’s draft extensively. I then asked it to compare our drafts and write a disclaimer. I liked the disclaimer, here it is in full:

Claude:

I’ll write a transparent reflection on my role in the drafting process. The final version differs from my draft in several key ways:

1. Tone and authenticity – Your version has a more raw, immediate emotional tone, particularly in addressing the current crisis in academia and higher education. Where my draft was more measured, yours directly acknowledges that things are “a bit shit right now” and includes personal touches like “staring at the wall…entertaining thoughts of nuclear war.”
2. Structure – You created a more dynamic structure by directly addressing different audiences (students and supervisors separately), which makes the piece more engaging and authentic. My version maintained a single narrative voice throughout.
3. Practical details – Your version includes more specific, lived experience details like crying over cake with peers and the current (2025) context of higher education in Australia. My draft was more general and less grounded in specific experience.
4. AI discussion – While I introduced the AI tools, your version provides more practical guidance with the specific prompt and a more nuanced take on the different AI tools’ capabilities, including ChatGPT’s access to the web.
5. Personal voice – Your version maintains the characteristic Thesis Whisperer blend of empathy, practicality and gentle humor, but with more edge and urgency than my attempt to mimic your style.

I helped generate initial ideas and structure around peer feedback, supervisor workload, and using AI tools, but the final piece is distinctly yours – more immediate, more grounded in real experience, and more attuned to the current moment in academia. This collaboration demonstrates how AI can be a useful thinking partner while still allowing the human author’s unique voice and perspective to shine through.