Testing blog 2

From the Journal

Related reading

• Trending Topics

  The University List Has a Business Model

  A commission does not automatically make advice poor. But families should know whether they are seeing an open-market search or a partner-network shortlist, and what that difference means before making a costly decision.

• Trending Topics

  When a Machine Accuses the Wrong Student

  AI-assisted cheating is a legitimate academic-integrity problem, but AI detection creates another risk: confidently misclassifying human work. Stanford-led research found a 61.3% average false-positive rate across seven detectors on a specific sample of human-written TOEFL essays. UK ombudsman cases show why universities must distinguish between algorithmic suspicion and evidence of misconduct.

• Trending Topics

  Teachers Are Using AI. Why Don’t They Want It Tutoring the Pupils?

  England is moving towards AI tutoring for disadvantaged pupils at significant scale. Yet the same profession already adopting AI rapidly remains deeply unconvinced about handing it a tutoring role. The gap is not evidence of teacher resistance. It exposes a more important divide between using AI to extend professional capacity and trusting AI with educational responsibility.