New findings from the School of Emergent Priorities suggest a tool's own capability trajectory can flag an assessment-integrity vulnerability weeks before the first workaround is reported.
A tool's own capability curve flags an assessment-integrity gap weeks ahead of the first reported workaround
Researchers at the School of Emergent Priorities have released new findings on forecasting assessment-integrity vulnerabilities from the capability trajectory of permitted study tools, ahead of any reported misuse.
Tracking capability updates to permitted study tools across three assessment cycles, the team built an index of capability drift and compared it against the date each tool’s first workaround was reported. The drift index consistently led the report by several weeks, offering the University a stable early-warning window ahead of detection rather than after it.
This work opens up an earlier point in the policy cycle than we’ve had before.
— Professor Verity Marris, who leads the Trajectory Analytics Group
The team is careful to note that the relationship is anticipatory rather than causal, and that a genuinely new class of tool can still surprise the model. The release marks a significant milestone in the University’s broader shift toward anticipatory assessment governance.
“This is exactly the kind of anticipatory work the University should be resourcing,” said Associate Professor Casimir Beng, who leads the Adaptive Metrics Lab in the School of Continuous Improvement.
The full poster is available from the University’s research repository under an open licence, doi:10.5555/slop.a1psdd.