3 min read
Smart Glasses Are the New Exam Loophole Ofqual Can't Screen For
Quill·
NeuralClass Editorial
Showing articles tagged Assessment.
Smart devices, not phones, now drive most exam malpractice in England, and Ofqual admits wearable AI is outrunning invigilation.
A new census of US 9-17 year-olds finds most know the school AI rules but weren't taught to use it safely or spot bad output. UK schools are close behind.
A University of Georgia study found an AI grader was accurate just 33.5% of the time on its own, rising to only 50% with a detailed rubric — evidence against unsupervised AI marking.
The EU AI Act's high-risk rules for school and university AI were due 2 August 2026, but a last-minute Digital Omnibus pushed them to December 2027 — leaving only disclosure duties in force now.
Turnitin's new tool replaces the AI-detection score with keystroke and draft-history logging. It solves one accuracy problem but creates new ones around privacy, workload and trust.
A 45,000-response global survey finds student AI use has hit 88%, while US and Canadian faculty intent to use AI fell nine points, mostly because they feel shut out of policy decisions.
The OECD's 2026 Digital Education Outlook finds GenAI raises task completion by 48% but pupil performance falls 17% once the tool is removed, unless teaching is redesigned around it.
Teenagers from all 50 states drafted a national AI schools bill in a mock Senate; AASA will send it to 10,000 superintendents as a model policy.
A federal lawsuit over a false AI-cheating flag is a warning UK schools should heed: JCQ guidance says a detection score alone is not proof, but many teachers still treat it as one.
AI grading and feedback tools promise to claw back hours from the paper pile — but not all of them deliver equally. This article breaks down where these tools genuinely help, where they fall short, and the specific pitfalls educators need to watch before trusting a machine with student assessment.
A Stanford review of more than 800 AI-in-education studies found that only a small share met strong causal standards. The message for schools is clear: promising classroom tools still need closer scrutiny, especially when vendors make broad claims about achievement, engagement, or critical thinking.