ACCA scraps remote exams as AI tools fuel global cheating surge

ACCA’s move is a case study in how fast high‑stakes assessment is buckling under everyday AI. In just a few years, remote proctoring tools that looked sophisticated during COVID have been overwhelmed by AI‑assisted answer engines, transparent overlays and covert earpieces. When the world’s biggest accounting body abandons online exams, it’s a signal that our institutional trust mechanisms aren’t keeping pace with cheap, accessible model capabilities.

This won’t change the frontier of model development, but it will shape the context those models grow into. If professions respond to AI by retreating to analog processes—pens, paper, in‑person invigilation—that’s a brake on some forms of digital transformation, even as back‑office workflows quietly automate. It also raises a deeper question: in a world where assistants can solve most textbook‑style questions, what exactly are we certifying when we certify a human?

Over the longer run, the episode pushes regulators and educators toward new verification paradigms—continuous assessment, practical simulations, even AI‑for‑AI monitoring. That arms race between detection and assistance is part of the broader AGI story: aligning not just the models, but the institutions around them, to preserve any meaningful signal of human competence.