Group assignments were already the most contested form of assessment, and generation has intensified the specific problem they were criticised for. The adaptations are visible in how these tasks are now set.

Why group work was already contested

The complaint predates any of this: effort distributes unevenly and the grade does not, so a shared mark rewards whoever contributed least.

Institutions kept group work anyway, because collaboration is a stated outcome of most programmes and cannot be assessed individually.

The defence rested on the assumption that a substantial artefact required substantial combined effort.

How generation changed the division of labour

A student can now produce a complete draft of the group's deliverable alone, in an evening, without consulting anyone.

That collapses the negotiation the assignment was designed to force, since there is no longer a reason to divide sections and reconcile them.

Groups that do this submit coherent work and learn nothing the task intended, which is harder to detect than uneven effort was. Uneven effort at least produced a document with visible seams, and a single-author draft has none.

Why contribution records became evidence

Shared documents with revision history, version control and project boards all record who did what and when.

Requiring the work to happen in such a system converts collaboration from a claim into a log, and the log is examinable.

The records are gameable with effort, though gaming them convincingly takes longer than participating would have, which is the useful property. Any verification method with that characteristic is worth more than one relying on detection after the fact.

How assessment split from the artefact

Many courses now grade the artefact lightly and the individual contribution heavily, reversing the previous weighting.

Individual components include a personal reflection on a specific decision, a defended section, or an account of a disagreement within the group and how it resolved.

These are hard to produce without having been present, which is the property the shared document lost. A student who missed every meeting can describe the deliverable and cannot describe the argument that shaped it.

Where the collaboration is real

Tasks that genuinely cannot be done alone survive the change without modification.

Those include work requiring simultaneous roles, data each member gathers separately, or a deliverable large enough that one person cannot review it in the time available.

Designing for that condition is more work than assigning a shared report, and it is the only version of group assessment where the collaboration remains the point rather than an assumption.