Still Yours: a doctoral researcher's guide to AI

A guide to using AI in postgraduate research, written with Benita Olivier.

AI can now produce a great deal of what a research degree asks for. The book takes that seriously rather than arguing around it: the thesis is the artefact, and the researcher is the point.

What it is

Still Yours: A Doctoral Researcher's Guide to AI — cover mockup, Benita Olivier and Michael Rowe

Still Yours: A Doctoral Researcher’s Guide to AI (Springer Nature, expected November 2026) is written with Benita Olivier, who works with postgraduate researchers across a wide range of disciplines.

Most of the available guidance sits at one of two extremes. At one end are detection, prohibition and institutional policy; at the other, tutorials for particular tools. We think that neither of these positions reaches the question a researcher actually has at their desk on a Tuesday evening, looking at a draft that isn’t working: should I use it for this, right now, and what happens to me if I do? That question cannot be answered in general, because the answer depends on what the task is for and on what the researcher is supposed to become by doing it.

The book starts from the question researchers are already asking themselves — why do a research degree at all, when AI can produce much of the work? — and treats it as an honest response to an honest situation rather than a failure of nerve. The answer it gives is that a research degree was never primarily about completing the tasks. The thesis is the proxy; the researcher is the point. AI can produce outputs, and it will keep getting better at producing them, but it cannot produce the researcher, and the capacities that a degree builds are the ones an AI-saturated world will need the most.

That answer suggests an uncomfortable implication, which the book treats as its central problem. Developing the expertise to judge whether an AI’s output is any good requires the very engagement that AI makes it easiest to skip. Reading closely, holding an argument in mind, sitting with a design until its weaknesses show; these practices are how taste and judgement develop, and they are exactly the parts of the work that are most tempting to delegate to increasingly capable AI models. So the practical question is not whether to use AI, but which parts of your work are building you, and what it costs to hand those over to a machine that’s only too willing to take it on.

How the book is organised

There are eleven chapters in three parts.

Part 1 is an orientation that covers what AI can actually do in research today, why a research degree still matters regardless, and how to work honestly and stay accountable for your thinking. The first chapter is about AI capability rather than specific tools: how these systems produce text, why this mechanism explains their characteristic failures, and why intuition can’t reliably find the boundary between what they do well and what they’re not good at.

Part 2, on working with AI, is the practical middle of the book, and it is organised by the kind of relationship a researcher has with AI rather than by stage of the research: AI as thinking partner, as research assistant, as writing collaborator, as mediator of the relationships around the work, and as project manager. We made this choice intentionally, as organising it by research stage would frame AI as a production tool and along the way, quietly settle the question the book is asking. Organising it by our relationship with AI keeps the developmental question in view, and also recognises that real projects rarely proceed in order.

Part 3 is about protecting your development and consolidating the argument: the skills you cannot afford to lose, how to demonstrate your own contribution, and what a research degree is for in a landscape that keeps moving.

Three ideas run through all of the book. The first is a way of reading any particular use of AI; whether it substitutes for work you would otherwise have done by hand, adapts how you engage with that work, or transforms what kind of work you can attempt. The second is a set of diagnostic questions to put to your own practice: does this build your sense of the field, do you know what to do next, and are you developing the capacity to direct work you could not have done alone? The third is accountability for the cognitive work; whether you can defend the decisions that shaped what is now in front of you, which is a harder test than whether you followed the rules, and one only you can answer.

Who it is for

The primary audience for the book is the doctoral researcher but it’s written for anyone doing supervised research that ends in a thesis: PhD and professional doctorate candidates, MPhil, Masters by research, MRes. Supervisors, research development staff and programme leads will find it useful too, particularly where a department wants to say something more specific to its students than “follow the policy”.

The book stays at the level of principles and worked practice, on the grounds that named products date within months while the reasoning about what to delegate to AI holds for longer, and is therefore the more sustainable lesson.

Status

Publication is expected in November 2026. The exact date is not yet confirmed.

Find out more

The book has a landing page at researchmasterminds.com/ai-and-your-doctorate, where you can read more about it and sign up to hear when it is published.

Related work on this site: