If one were to bring a product to market that plays Russian Roulette with qualitative & quantitative Validity & Reliability (cf. Cohen & Morrison 2018:249-… Table 14.1 & 14.2), would it then be surprising that their “paper,” justifying that same product, plays it loosely as well? (Ref: https://lnkd.in/ek8nRxcF and https://lnkd.in/e5sGtMSH )
…and, before one rebuttals, does a technical paper (or, “report”) with the flair of an academic paper not need to follow attributes of validity & reliability? Open(AI)ness / transparency are but two attributes within an entire set, innate to proper scientific methodologies AND engineering methodologies; the 2 sets of methodologies not being synonymous.
Open is lost in AI.
Or, as argued elsewhere, as a rebuttal to similar concerns with LLMs, are these “only for entertainment?” I say, expensive entertainment; i.e.: financially, socially, environmentally and governance-wise, not to mention the expense on legality of data sourcing. (thread: https://lnkd.in/db_JdQCw referencing: https://garymarcus.substack.com/p/what-did-they-know-and-when-did-they?r=drb4o and https://www.animasuri.com/iOi/?p=4442 )
“Ah, yes, of course,” one might reply, “these are the sad collateral damages as part of the acceleration risks, innate to innovation: fast, furious, and creatively destructive. A Luddite would not understand as is clear from your types and your fear-mongering during them days of the printing press.”
no. Not exactly.
Reflection, care, (re)consideration, and nuancing are not innate to burning technology at the stake. Very much the opposite. Love for engineering and for sciences do not have to conflict with each other nor with love for life, ethics, aesthetics, love for relations, poetics, communities, well-being, advancement and market exploration.
Rest assured one can work in tech, play with tech and build tech while still reflecting on confounding variables, processes, collateral affects, risk, redundancies, opportunity, creativity, impact, human relation, market excitement and, so on.
Some humans can even consider conflicting arguments at once without having to dismiss one over the other (cf.: https://t.co/D3qSgqmlSf ). This is not only the privilege of quantum particles. Though, carelessness while seeing if, when and where the rouletted bullet lands, while scaling this game at global level and into the wild, is surely not a telltale signaling of such ability either.
Are some of our commercial authoritative collectives too big to be failed at this? No testing ability in the highest percentile answers us this question folks. That “assessment” lies elsewhere.
And yet, fascinating are these technological tools.
—-•
Additional Triggers:
Thank you Professor Gary Marcus
Thank you Sharon Goldman
—-•
Cohen, L., Manion, L., Morrison, K. (2018). Research Methods in Education. 8th Edition. Routledge.
—-•