Artificial Intelligence (AI) is here, and it’s transforming how we design learning experiences. For learning designers, AI offers exciting opportunities to streamline workflows, enhance personalisation, and improve accessibility but these benefits depend on how outputs are verified and whether they remain valid under real conditions. Additionally, with these benefits come important ethical considerations. (If you’re exploring how these decisions about efficiency, scope, and ethics sit within the wider responsibilities of the profession, the overview in The Role of a Learning Designer sets out the expectations designers balance when adopting new tools.)
In this post, we’ll explore practical ways to use AI tools today and reflect on the responsibilities that come with them.
Generative AI for Content Creation
One of the most immediate applications of AI in learning design is content generation. Tools like ChatGPT, Claude, and Gemini can help create draft text for eLearning modules, discussion prompts, or even assessment questions. For example, when developing my latest course, I used the university’s enterprise version of CoPilot to produce scenario-based examples illustrating concepts like cognitive biases and motivation, to bounce ideas off to create skills-based activities, to source the references for quotes, and source text-based alternatives to videos in case learners can’t access them for any reason. These examples provided a starting point, which I then refined and verified, because AI can assist with generation but not judgement, but it did save a lot of time in a time poor course production cycle.
Generative AI can also assist with multimedia creation. Platforms such as Synthesia or D-ID allow designers to produce video content with AI-generated avatars, reducing production time and cost. Some of the videos I seen generated are incredibly impressive, much more so than the Coca Cola ad this year. However, AI outputs should always be reviewed for quality, tone, inclusivity, and factual correctness, as they may appear coherent without guaranteeing accuracy or pedagogical validity. (If you’re pairing AI‑generated scripts, visuals, or narration with multimedia elements, the guidance in Top Multimedia Principles for eLearning Design helps ensure the final content supports cognitive processing rather than overwhelming learners.)
Personalisation and Adaptive Learning
AI-powered personalisation is another potential game-changer. Adaptive learning platforms like Area9 use algorithms to tailor content based on learner performance and preferences. Theoretically, each learner receives targeted support where they need it most, creating a more efficient and engaging experience, If the underlying pedagogy is right, though these systems constrain what can be inferred about learners and shape what becomes visible to designers. For personalised learning to be fit for purpose, adaptive logic needs to align with pedagogical principles rather than purely algorithmic efficiency. The system needs to be designed for human learning and not technological.
The scalability of personalised learning like this puts a lot of pressure on learning designers to get it right. I’ve experimented with this on a smaller scale in the form of branching scenario tasks and it creates a lot of moving parts even on a small scale. To design content that an AI system can reorganize dynamically add even more layers of complexity. (When working with adaptive systems, it’s also essential to evaluate whether the personalisation actually improves learning. The frameworks in Measuring the Impact of Learning Programs offer practical ways to assess what AI‑driven adaptations are achieving.)
Improving Accessibility with AI
Accessibility is dear to my heart and a core responsibility for learning designers, and AI can help us meet it more effectively, but accessibility remains a design condition, not something tools can solve automatically. Tools like Microsoft’s Immersive Reader or AI-driven captioning services can make content more inclusive for learners with diverse needs. For instance, automatic transcription and real-time translation can support multilingual audiences and I’ve taught team members how to AI-generated alt text to improve screen reader compatibility.
We have started putting AI to use to generate descriptive alt text for complex diagrams. It’s sometimes difficult to know where to start so AI solves that particular difficulty, making WCAG compliance more likely and helps save a lot of time though each description is still manually checked for accuracy and AI hallucinations. (If you’re strengthening your overall accessibility practice alongside AI‑supported alt text and transcription, A Designer’s Guide to WCAG Perceivable Principles offers practical checks that pair well with AI‑enabled workflows.)
Ethical Considerations: Accuracy, Bias, and Transparency
AI is exciting and offers a lot of potential for improving course creation as well as learner experience of a course but there are a lot of ethical questions surrounding its use.
Environmental concerns like water and power consumption, copyright concerns around what the AI is trained on, and data privacy and security concerns around who and what has access to what we in put into AI and why are all worries before looking at the actual generated content which is only as good as the material it is trained on. Tay where a Microsoft chat bot for then Twitter became a racist holocaust denier in a few days is an early example of how those materials are key, but even now underrepresentation of women and global majorities in data sets as well as preexisting prejudices within the data (racism, sexism, agism, homophobia, etc in texts, posts, articles, literature and other training materials) skew the outputs.
I’ve used AI to generate images for this blog and got so annoyed at it only giving me men, usually white, able bodied, and sometimes (weirdly) balding even when I explicitly asked for diversity, that I gave up and use line art of objects instead. Designers must verify all outputs against reliable sources, and we need to tell learners when AI has been used in creating their learning materials so they can check it too. (These questions also sit within the broader conversation about how AI is reshaping learning design, explored from a future‑focused angle in The Future of AI and Machine Learning in Education.)
AI can generate content and tasks that mimic creativity, empathy, and critical thinking but it is no substitute for the real thing. In fact, one of my favourite tasks to give learners at the moment is for them to interrogate AI outputs in one form or another to develop those critical thinking skills that we all need when interacting with AI. I might start throwing in some AI ethical conundrums into those tasks too to really get the brain cells moving because AI is really, really useful and it isn’t going away so we all need to get really really good at upholding ethical standards and thinking critically.
Leave a Reply