Why educators should consider this toolkit

Students are already using AI

Students are already using AI. The real question is no longer whether they will use tools such as ChatGPT, Claude, Gemini or Copilot, but how they will use them.

Independent student AI use is often unstructured. A student may paste in an essay and ask directly for it to be fixed: “Improve this”, “Make this sound more academic”, “Fix my writing” or “Rewrite this”.

Even students who are more cautious with their prompts are using systems that are designed to be helpful. The AI may jump in, without being asked, and fix the work for them.

Educators are now seeing the results: ghost-writing, over-polishing, dependency and superficial improvement. Increasingly, reading this kind of over-polished work can become exhausting and dispiriting because the student’s own voice starts to disappear.

The AI Personal Tutor Toolkit offers a different approach. It is designed to help students use general-purpose AI as a focused writing tutor, not as a replacement writer.

It gives students a way to use self-imposed guardrails, but the positive case is just as important: structured and specialist writing support that is more closely aligned to academic development than a general AI chat normally provides.

The gap it fills

Many existing AI education products are institutional platforms. They are built for schools, colleges and universities that want managed accounts, teacher dashboards, integrations, assignment workflows, analytics, compliance features and, in some cases, academic-integrity monitoring.

Those systems can be useful, especially inside managed lessons or formal assessment workflows. But they can be costly, need technical integration, and may not address the independent study that happens outside formal platforms. In universities especially, much student learning happens privately: planning, drafting, revising, reading, studying and preparing for feedback conversations.

At the other end of the spectrum is raw LLM use. Students already have access to powerful AI systems, but those systems are open-ended. They will often do whatever the student asks, including rewriting, simplifying, summarising, polishing or producing answers. Their default is often to leap in and solve the problem for the user.

The Toolkit provides a lightweight, low-cost and inspectable layer of structure that can guide student and teacher use of general-purpose AI tools.

What the Toolkit is

The AI Personal Tutor Toolkit is a set of prompt libraries that guide an AI system to behave more like a personal tutor.

Instead of asking the AI to produce polished answers, the prompts ask it to diagnose problems in a student’s own work, explain mistakes clearly, ask questions, give targeted practice, help students revise, turn feedback into action plans, support academic thinking, preserve student authorship, and encourage honest AI-use records.

Students can still use AI, but the AI should support learning rather than replace the learner.

Why this matters for teachers and lecturers

Educators face a practical dilemma. They may not want to ban AI completely, but they also do not want students to outsource thinking, writing or assessment tasks.

The Toolkit gives educators a way to say: “You may use AI, but use it in this structured way.”

Instead of only asking, “Did the student use AI?”, educators can ask: “Did the student use AI to understand feedback, improve their own draft, practise a skill or reflect on their learning?”

That is a much more constructive position.

Why universities should care

Universities are in a particularly difficult position. They can provide approved tools, policies and guidance, but they cannot realistically control every student’s independent AI use. Students will use AI in their own time, on their own devices and for their own study processes.

That means universities need more than detection and compliance. They need usable models of responsible independent AI study.

The Toolkit is interesting because it offers exactly that: a structured way for students to use AI outside formal classroom systems while still following educational principles.

It could support study skills provision, academic writing support, learning development teams, dissertation and research proposal support, transition-to-university programmes, academic integrity guidance, AI literacy teaching, and formative feedback workflows.

It is not a replacement for university policy. It is a practical bridge between policy and student behaviour.

The cost argument

A major advantage is that the Toolkit does not require a new platform.

Institutions already face an expanding market of AI tools, many of which come with licence costs, procurement processes, staff training demands and long-term vendor dependency. The Toolkit works differently.

It is closer to an open educational resource: a structured set of prompts that can sit on top of existing AI access. That means the main ongoing cost is whatever AI access the institution, teacher or student already uses. The Toolkit itself can remain portable, adaptable and inspectable.

This matters for schools, colleges and universities that want to experiment with AI-supported learning but are not ready to commit to another commercial platform.

Writing is thinking

The educational case is not only about preventing misuse. Writing is one of the ways students think. When they struggle to form a claim, connect evidence, order a paragraph or explain a concept, they are developing understanding.

If AI takes over those choices too early, students may receive smoother text while losing the concentrated thinking the task was meant to develop. The toolkit is designed to support that thinking with focused feedback rather than replace it with finished prose.

The pedagogical argument

The strongest reason to look at the Toolkit is not cost. It is pedagogy.

Raw AI tools tend to optimise for helpful completion. If a student asks what the AI thinks of a paragraph, the AI may simply produce its fixed version. That can be useful, but it can also bypass the learning process.

The Toolkit tries to slow that process down. It encourages a cycle of diagnosis, explanation, practice, revision and reflection.

That is much closer to formative teaching.

For example, instead of simply rewriting a weak paragraph, a structured tutor prompt can ask the AI to identify the main issue, explain why it matters, show the student what kind of move is needed, and help them revise the paragraph themselves.

This is the crucial distinction: the aim is not better AI-generated text. The aim is better student decision-making.

The AI-literacy argument

There is a second pedagogical gain beyond protecting any single assignment. Students who use the toolkit are not only getting better feedback; they are learning how to direct an AI tool responsibly.

Unstructured AI use can teach one lesson: the machine produces the work. Structured use teaches a different and more durable set of habits — diagnose before fixing, separate meaning from wording, ask for explanation rather than answers, and stay the decision-maker.

For institutions writing AI-literacy outcomes into their curricula, this is useful. The toolkit is not a guarantee of good AI use, but it is a concrete method for showing what responsible, learning-focused AI use can look like in practice.

How educators might use it

A lecturer might recommend it to students as a responsible way to get feedback on early drafts.

A learning developer might use it in workshops on academic writing, critical thinking or revision planning.

A schoolteacher might adapt the prompts for GCSE or A-level writing practice.

A dissertation supervisor might use the research proposal prompts to help students test the clarity, feasibility and logic of a project.

A university might include it in AI literacy guidance as an example of acceptable, learning-focused AI use.

A department might adapt the prompt libraries into its own discipline-specific versions.

Why it is not just another prompt collection

The Toolkit is more interesting than a list of prompts because it treats prompts as educational infrastructure.

It brings together organised prompt libraries, a teaching approach, tool menus, prompt chains, testing and audit materials, versioning, academic-integrity boundaries, and guidance around tutor-like behaviour.

That makes it more like a small pedagogical system than a random set of tips.

For educators, this matters because the approach is inspectable. They can see what the prompts are asking the AI to do. They can adapt them. They can test them. They can discuss them with students.

That is very different from telling students to “use AI responsibly” without giving them a concrete method.

Evidence and evaluation

This is a design argument, not yet a claim of proven large-scale impact. The case for the toolkit is based on its structure, teaching principles, audit process and early use. Formal evaluation would be needed to make stronger claims about outcomes, time-saving or learning gains.