RE:
https://scholar.social/@olivia/116895347041158720
"When technology companies offer teachers and students ‘AI training’ or promote ‘AI literacy,’ they are engaged in brand marketing, framed as a progressive, philanthropically fundable intervention.
The literacy label locates the problem in users’ understanding rather than in systems’ design, and in doing so, displaces accountability from the companies that built and profit from these systems toward the educators and students who are expected to become fluent in them, and forecloses the prior question of whether fluency is the goal at all.
[...]
The [stochastic parrot] dismissal reflex stabilizes prematurely into the metaphor. The [AI literacy] accommodation reflex senses that the parrot metaphor has become a comfort blanket and concludes that the critique must be updated rather than sharpened.
Without a clear account of what statistical pattern-matching can and cannot do, ‘AI literacy’ remains undefined, critical pedagogy cannot say what it is protecting, and analyses of labor displacement cannot specify what is being displaced.
The stochastic parrot formulation, properly understood, does not function as a conclusion but as a constraint on interpretation; these system produce outputs that resemble plausible continuations of their training distribution, without grounded inference, semantically anchored representation, or the kind of iterative, error-sensitive engagement with the world that characterizes thinking. These are not limitations that scale resolves; they follow from the mechanism itself."
Open quoted post Quoting
New preprint! So honoured to have worked with Ishani Ray on providing vital critique on "AI literacy" & correctives for those interested in understanding why "just use the newest version of an LLM" is nonsense
Contra Literacy-Laundering: Mechanistic Critical AI Literacy
https://doi.org/10.5281/zenodo.21222978
1/
