Blinded by the (traffic) lights: The intellectual bankruptcy of AI use scales

https://www.linkedin.com/pulse/blinded-traffic-lights-intellectual-bankruptcy-ai-use-mark-a-bassett-0davc/

“AI use scales are taking up space that legitimate responses to the challenges of generative AI need to occupy, and they are increasingly defined by negation. Dismissing valid critiques of these frameworks as ‘oversimplification’ and ‘misinterpretation’ has become the standard defence, a manuevre that allows such frameworks to endlessly absorb scrutiny without ever being held to a fixed claim about what they actually do. The enforcement gap is waved away as someone else’s implementation issue. The incoherence of the middle bands, where a single piece of student work can simultaneously satisfy and violate adjacent levels, is reframed as productive flexibility. The unresolvable question of when an assessment begins is set aside as a philosophical curiosity. The fact that generative AI systems make no distinction between activities that one level permits and those that the next level prohibits is entirely ignored. The diffusion of AI capabilities into the everyday software environments students already inhabit, which renders the very idea of switching AI on and off for a task incoherent, is not addressed. The harm done to students who follow institutional guidance while their peers do not is left for the reader to discover. What remains is a label on a task that communicates an institutional preference about conditions no institution actually controls, propped up by a defence that grows thinner with each revision.”