From a presentation given at the e-AA International Conference 2026 by James Spurgeon, Strategic Architect, RM and Rita Bateson, Co-founder and Head of Education, Eblana Learning, based on the presentation slides delivered and the discussions they prompted.
Most general and professional qualifications are a solitary affair. A single candidate tackling the set tasks on their own to demonstrate their own proficiency. However, in many cases it is not a realistic representation of the normal environment where the candidate will apply the learning being assessed. For example, much workplace activity involves working as a team and its value gets credited accordingly. Individual assessments prevail in subjects like music performance, but when was the last time you witnessed the possibly slightly forlorn sight of a lone cello, viola or clarinet recital?
Assessment of groups as a means to establish competence is still rare, actively avoided even. Among the reasons for avoiding group assessment are:
Validity and reliability concerns centre on the ability of group assessment to accurately demonstrate which of the group actually learned what in any given scenario. Group dynamics can negatively impact the assessment process when some group members dominate proceedings. Others remain silent, in some cases actively choosing to ride on their colleagues’ coat tails making assessment of their learning almost impossible.
Groups inevitably comprise candidates with different characteristics. There will be introverts, as well as extroverts. Some may not be native speakers of the language of the assessment or have cultural norms which inhibit their full participation. These elements all create concerns about the equity of group assessment.
Practicality concerns revolve around time, marking and standardisation. It can be difficult for remote markers especially to determine which candidate is which. As a result, linking each learner to the contribution they make presents problems. Logistical hurdles present themselves in large-scale group assessments. Moderation of remote marking without individual scripts to refer to becomes tricky. Individuals in the group could benefit from prompting by colleagues that is not evident to markers, or even being replaced by proxies.
Group assessment also potentially creates confusion in terms of what constitutes a contribution worthy of being rewarded with marks. In a group environment, does a specific positive contribution about the discipline being assessed deserve more recognition than something that merely keeps the group endeavour on track for everyone?
RM is at the forefront of developing AI for marking and feedback, mostly focused on marking assessment artefacts created by individual candidates, not groups. However, rapid development of AI technologies has created tools that show a potential route for overcoming some of the challenges group assessment faces. Parsing recordings or transcripts with AI tools can reveal where and when interruptions happened and where the power dynamics lay.
In situations that value the expression of ideas, knowing who originated the idea as opposed to who actually voiced it can be invaluable. AI tools can help determine the source of an idea, the identity of anyone who subsequently developed it and who merely repeated what others verbalised earlier.
By effectively using prompts (that could be formulated in advance of assessments being marked) it would be possible to classify each speaker into categories such as quiet expert, over talker, facilitator, decision maker, or observer. This would help markers assign marks to individual candidates with confidence.
If group assessment becomes a credible and trustworthy option, what new skills or knowledge will it become possible to assess? What new questions we could ask?
AI tools have the potential to increase the viability of group assessment. Currently the tools that suggest this possibility are not specifically designed with general or professional qualification assessment in mind. However, they do point the way to a development path that could deliver suitable tools in the future.
The EU AI Act is a potential hurdle to negotiate, since it categorises AI use in systems “intended to evaluate learning outcomes” as high risk. However the Act does not seek to forbid their use for such purposes, rather to ensure its safety for all stakeholders. For this reason, perhaps professional qualifications will be where developments happen first.
Find out more about Rita’s work on AI with schools and other educational institutions around the world at the Eblana Learning website.
Header image courtesy of Telling Photography