Beyond Cheating: Does the Current Assessment Model Still Work?
4 min read
Assessment & Integrity
Newsletter 6 of 14 in the WEF series. Turnitin's 2025 survey found 95% of educators believe AI is being misused. But the bigger question isn't about cheating — it's whether the assessment model itself still works.
If a learner can submit an impressive answer but cannot explain, defend or apply it without the machine, the assessment is no longer measuring what it claims to. Part 3 of the Educational Futures series.
The OECD has just published the first draft of a test that will, for the first time, put a number on how well 15-year-olds can navigate AI and media – PISA 2029's Media and Artificial Intelligence Literacy (MAIL) framework.
When AI can produce the essay in seconds, the essay can no longer be the evidence. The question isn't how to detect AI-written work – it's what we're actually measuring when the artefact is no longer the point.
Paul Penfold is an education and technology consultant based in Chiang Mai, Thailand, with six decades of experience spanning classroom teaching, academic research, government policy, and humanitarian service. He advises schools, ministries, and EdTech companies on AI in education, alternative education, and learning technology, and has published 37 papers on immersive and virtual-world learning. His career includes refugee education on the Thai-Cambodian border, pioneering virtual-world hospitality education at Hong Kong Polytechnic University, and national tourism HR strategies across Asia and the Caribbean.