•
“Democratization of AI”

This phrase was used by a leading industrialist in a 2016 interview[1] It was a talk between 2 individuals. The 2 males used a number of concepts from the field of ethics. “Good” being 1

1 spearheads what is now a less-open approach to a less-transparent technology. A Paywall is enacted. The full version’s accessible to those who have the money per month, & who live in specific regions of the world. This is while many in that realm are having the preexisting resources enabling to check whether the tech’s output’s sufficiently valid & credible

It could be argued that in early 2023 democratization of the tech has already occurred for those who do not urgently need democratization, & yet who are enabled to be concerned about access. That includes me

Let us now consider “democratization of knowledge”

Those initiatives preceding 2016, that have been promoting democratization of knowledge, eg through Open Access, Creative Commons, decentralization of transparent editorial oversight, relatable legal constructs in support of these, or via far more radical means, could be aided by “democratization of AI.”If it too were open & if the threshold of onboarding is manageable. Is it sufficiently so?

Some radical initiatives in this democratization process are labeled “piracy”

Where do the datasets of LLMs come from?

Pirate’s a label invented by gatekeepers. These control workers, some of whom have to pay to deliver their work, presented as increases to collective knowledge, secured behind paywalls. The latter is a number of academics, published in what are perceived as respectable journals & which are, by some, considered a reason for increased tuition for access to universities

Democratization of AI is not the same as democratization of knowledge. Democratization of simulacra is not the same as democratization of AI, nor of access to references. Democratization of industrial access to data, generated by any individual, is not democratization

The tone of authority from those who output the narrow type of AI (while hinting at neurosymbolic AI), the tone of their huge models, investments & budgets, the authoritative tone of the output of their bots, the irresistible accelerating drive of not missing out —all this, irrespective of validity or credibility of the information from the stakeholders, as well as their bots— these do not equate to “democratization of knowledge”

The proliferation of anyone’s data anywhere, at any time, for any purpose, sounds like a treasure throve for those in favor of democratization of knowledge

Who are the “demos” in democratization?

We the People. We, People. We love good? We love open? We the People, are we the people anywhere, anytime, all at once?

The tech is “good.” The tech has open access to your data

We, some Platonic People, are market segmentations, resources, & resource allocations. We are the tech’s monetization. So, I pay.

#ailiteracy #aiethic

reference

[1] https://www.youtube.com/watch?v=tnBQmEqBCY0&t=1s (“democratization of AI” at 11:53)