Personalisation is difficult to criticise.
Giving learners more control over their experience appears self-evidently positive. Greater flexibility, more pathways, adaptive content, optional resources, and learner-directed journeys all sit comfortably alongside contemporary discussions about inclusion, autonomy, and learner-centred design.
When a project brief asks for personalised learning, few people immediately ask whether there might be such a thing as too much choice. Yet the question is worth examining. Personalisation is often discussed as though it exists on a simple scale where more choice produces better experiences. In practice, every decision about choice introduces trade-offs and the challenge is deciding how much choice is appropriate, where those choices should sit, and who carries the consequences when they fail to support learning rather than whether learners should have choices.
This is a design problem, not a technology problem, and, as with many design problems, several defensible solutions exist.
Option One: Maximise Learner Choice
One response is to provide as much autonomy as possible. Learners choose their preferred formats, learning pathways, activities, assessment routes, pace, sequence, and supporting resources. The designer’s role is to create opportunities and allow learners to determine what works best for them.
The attraction is obvious. People differ. Learners arrive with different experiences, goals, preferences, responsibilities, confidence levels, and constraints. More choice appears to acknowledge this variability rather than forcing everyone through a single route.
A parent studying around caring responsibilities may value flexibility differently from a full-time student. An experienced professional may benefit from bypassing content that a novice needs to complete. Greater autonomy can support agency, ownership, and self-direction.
The difficulty is that choice has costs.
Every additional decision creates additional cognitive work. Learners must evaluate options, predict consequences, compare alternatives, and determine which route best serves their needs. This burden is not distributed equally. Confident and experienced learners may navigate complex choices relatively easily, while others become uncertain, overwhelmed, or disengaged. What appears empowering to one learner can feel unsupported to another.
At a programme level, extensive choice can also produce highly uneven experiences. Learners may complete the same course while encountering substantially different content, activities, or opportunities for practice.
The ethical risk is that autonomy becomes the priority while responsibility for navigating complexity shifts increasingly onto learners.
Option Two: Provide Guided Pathways
A second response is to reduce the number of decisions learners are required to make. The core pathway is clearly structured and intentionally sequenced. Personalisation exists, but within defined limits. Learners may choose between a small number of routes, access optional extensions, or select additional resources, while the overall experience remains guided.
This approach prioritises clarity. Instead of asking learners to determine the most appropriate pathway from innumerable possibilities, the design performs some of that work on their behalf. Navigational effort is reduced, expectations are clearer, and learners can focus more attention on learning itself.
There are strong accessibility arguments for this position. Clear pathways reduce uncertainty, support orientation, and can make participation easier under real conditions where learners may already be managing competing demands on their attention and time. The risk lies elsewhere.
Every guided pathway reflects assumptions made by somebody. Choices about what is required, what is optional, what comes first, and what can be omitted are design judgements. Those judgements may be defensible, but they remain assumptions.
A pathway intended to support learners can also constrain them. Opportunities for exploration may narrow. Alternative approaches may become less visible. Some learners may find that the pathway does not align with their circumstances, prior knowledge, or goals.
The ethical risk here is different from the first option. Rather than transferring responsibility to learners, it concentrates responsibility within the design itself. The question becomes whether the designer’s assumptions are justified.
Option Three: Use Data-Driven Recommendations
A third response attempts to balance autonomy and guidance. Learners retain choice, but recommendations are provided based on previous activity, learner characteristics, performance patterns, or system-generated predictions. Rather than requiring learners to evaluate every option independently, the system suggests what may be most useful.
This approach promises responsive support at scale.
Recommendations can reduce navigational burden while preserving flexibility. Learners remain free to make their own decisions, but they receive guidance informed by available data. For organisations, this can appear particularly attractive. It allows large numbers of learners to receive differentiated support without requiring continuous intervention from educators or support staff.
The problem is that recommendations are never neutral. Every recommendation system embodies assumptions about what success looks like, which indicators matter, what data should be collected, and how learners’ future needs can be inferred from past behaviour. The logic may be invisible to learners. The underlying data may be incomplete. The recommendations may reinforce existing patterns rather than challenge them.
Questions also arise about privacy, transparency, and accountability. If a learner follows a recommendation that proves unhelpful, who is responsible? The learner? The designer? The institution? The system itself?
The ethical risk is not simply that recommendations might be wrong. It is that their assumptions become difficult to see.
Which Option Is Best?
The difficulty with this problem is that none of these responses is obviously unreasonable. Each reflects a legitimate educational value.
- Maximum choice prioritises autonomy.
- Guided pathways prioritise clarity.
- Data-driven recommendations prioritise responsive support.
Each can be justified. Each can benefit learners. Each also introduces different risks.
- Zero choice may unnecessarily restrict learner agency.
- Unlimited pathways may transfer excessive responsibility onto learners.
- Recommendation systems may obscure the assumptions shaping learner experiences.
This is why the problem cannot be solved simply by adopting a principle such as “more choice is better” or “guidance is always necessary”. Real learners encounter real constraints. What supports one learner may create difficulties for another. The challenge is not eliminating trade-offs. It is recognising them.
The most defensible response is often neither unrestricted autonomy nor tightly controlled pathways. Instead, it involves creating meaningful choices while reducing unnecessary burden. Which option is most appropriate depends partly on the consequences of getting the decision wrong. Asking learners to choose between supplementary resources carries very different risks from asking them to choose assessment routes, progression pathways, or core learning experiences.
Learners need opportunities to exercise judgement, pursue relevant routes, and adapt learning to their circumstances. They do not necessarily benefit from carrying every navigational decision themselves. Equally, they should not be guided so heavily that autonomy becomes largely symbolic.
This is where personalisation becomes an ethical question rather than simply a technical one.
The issue is not whether learners should have choices. The professional responsibility is deciding which choices genuinely support learning, which choices create avoidable burden, and who is expected to carry the consequences when those decisions do not hold up under real conditions.
Leave a Reply