Three levels of AI proficiency for university educators

Many academics use generative AI (GenAI) to draft emails, summarise readings and brainstorm ideas for teaching and research. While these are valid uses, they require surface-level engagement with GenAI tools. As AI capabilities grow, so too do opportunities to use them to develop meaningful learning opportunities for students.

The above tasks belong to the first of three AI proficiency levels for educators. Here, GenAI is used to do what its name so effectively implies: generate answers.

At the second level, it becomes a collaborative partner in everyday teaching. Moving to the third level requires an educator to create structured systems and processes that put a GenAI tool to work on their behalf.

From asking to collaborating

At level one, educators can use GenAI to improve efficiency but not to fundamentally change the way they work. To move to level two, they must start actively collaborating with GenAI tools. In practice, this means using AI as a co-designer and thinking partner to help them test ideas, challenge their assumptions and refine their explanations.

At level two, instead of asking a tool to simply produce a lecture, educators use it to explore teaching approaches, identify likely student misconceptions and suggest interactive learning activities. Used this way, AI tools not only improve efficiency but also strengthen an academic’s clarity of thought.

Effective AI use at this stage depends less on the tool’s sophistication and more on the quality of the questions and prompts we feed it.

• Spotlight guide: Boosting AI literacy across your institution
• How to choose the right AI tools for teaching
• Only a human will do: when to eschew AI in teaching and research

From collaborating to delegating

Moving to level three proficiency means moving from one-off interactions to building small, reusable systems to handle routine tasks. This starts with the identification of recurring tasks that follow predictable patterns, such as writing feedback, summarising modules or responding to common student queries.

For example, an educator marking work for large cohorts might create a structured prompt linked to the module learning outcomes and marking rubrics. Instead of writing each comment from scratch, they can ask GenAI to generate personalised formative feedback based on key strengths, areas for improvement and advice. The educator still reviews and refines the output, but the process becomes significantly faster and more consistent. By developing structured prompts and templates in this way, academics can produce high-quality outputs without having to start from scratch each time.

Over time, these repeated processes evolve into workflows that reduce both time and cognitive effort. I like to think of these as academic “playbooks”. Each playbook defines a task, the required inputs, the steps to follow and the expected output. This approach helps educators move from ad hoc AI use towards more intentional and strategic integration into academic work. It is also more likely to prompt AI output that is aligned with academic standards and course objectives.

AI does not remove the need for expertise at this stage; it amplifies it. Clear instructions, well-defined assessment criteria and strong disciplinary understanding are fundamental.

How staff AI proficiency can affect academic outcomes

Many students are using AI at a basic level. If academics remain at level one, they are unlikely to demonstrate the skills and knowledge required to prepare students for the realities of an AI-driven world. By improving AI proficiency, academics can not only model effective use, but they can design learning and assessment that builds judgement and creativity rather than the ability to recall facts.

How institutions can help improve educators’ AI literacy

To progress towards digital transformation, universities must align institutional strategy with the realities of academic work.

Academics need time to experiment with GenAI, opportunities to share practice and access to professional development opportunities that go beyond basic training. This cannot happen without an institution’s support.

Leaders must also recognise AI capability as part of core academic practice. This means integrating it into teaching development frameworks and workload models. Without this, GenAI use is likely to remain uneven, with some academics advancing rapidly while others struggle to move beyond basic applications.

To become proficient AI users, academics must shift their approaches, starting with a single task, developing a simple workflow and then refining it over time. In a rapidly changing educational landscape, the ability to make AI work for you will shape not only academic productivity but also the kinds of learning experiences we create for our students.

Patrice Seuwou is associate professor of learning and teaching and director of the Centre for the Advancement of Racial Equality at the University of Northampton.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.