One striking aspect of the industrial-scale plagiarism that is current "AI", is that the mechanism of generation leads inevitably to soulless mediocrity in its output.
This is because it samples, selects, blends, extrapolates from multiple sources. It is based on averaging - which entails removing the extremes that characterize genuine creativity, including genius.
Consider: how would you plagiarize poetry such as to generate a work of genius?
The only valid method is to re-label.
You would need already to know of, a poetic work of genius (something which depends on prior human evaluation) perhaps a particular sonnet by Shakespeare.
...Then to re-label this sonnet with your own name as author instead of Shakespeare - and convince other people that you (not Shakespeare) had written it.
Such is the only way to ensure the production of a work of genius, of quality equal to the original.
If instead you were to try and create a poem by sampling and combining several or many Shakespeare sonnets, then obviously the result would not be as good.
If you tried to create poetry by (in some way, any way) averaging multiple Elizabethan poems of multiple authors, or many poems from all eras - the result would become less and less good, the more you added.
The more you sample, and the more you average - the worse it gets.
What this kind of multiplying and averaging can do, is create fakes that are hard to distinguish from mediocre poems - and can pass as an instance of the type, especially among people who are lack time, or aptitude, have little interest, or are weakly motivated.
For example, if you were to combine all of Shakespeare's 154 sonnets - you could perhaps make something that was not easily distinguishable in quality from the mediocre or poor instances of his Sonnets - especially among those who were not really able to appreciate the quality of the best of the Shakespeare originals.
And the AI approach is what successful forgers have done in the world of fine art. They do not try to forge the best Vermeer; not least because the forger could not do it. Instead, they attempt to forge a mediocre instance Vermeer, a variation on a known theme; something "good enough" that - when provided with a fake provenance by aesthetically-incompetent evaluators - can pass as a mediocre work by a great painter.
Such mediocre forgeries can then sometimes be passed off as "great art" on the basis that they are superficially appealing to only-mildly-interested people with mainstream modern tastes - but that does not affect their artistic mediocrity - evident to those who can evaluate quality.
But here we have another problem of AI, which is that it is mostly being casually used by lazy and conformist people who have little interest, ability or experience in the area of implementation.
If all you are working at some chore, and all you want is background music, then genuine creativity is irrelevant - fakes and forgeries are fine. Computer-generated or AI muzak will suffice.
If you are an average school kid, college student, post-graduate, professor; and all you want is to complete some assignment with the minimum of effort and getting the highest evaluation possible (without activating the plagiarism detectors - mechanical or human); then the truth and validity of what you write is irrelevant. Consensus-compatibility is very important, but ultimate value is not at all. Averaged mediocrity is indeed the ideal! AI will suffice.
If you are a professional researcher in "science" or academia, who seeks career advancement and high status; then you will do or say whatever you believe will help pursue those goals. Developing, buying, using or praising AI will be embraced insofar as it seems expedient in terms of your immediate objectives. The fact that AI-output is boring, derivative and adds nothing substantive is all-but irrelevant; because that exactly characterizes the professional environment in which you already operate and in which you hope to thrive. AI will be fine for your purposes.
It is evident that the set-up of current so-called AI - to sample (i.e. plagiarize) massively, combine and blend multiples, and generate an averaged-output; is necessarily soulless and mediocre.
AI simply cannot-help but generate the soulless and mediocre - and is therefore a tool introduced and imposed by those who desire that society should be like that.
Furthermore, this kind of "AI" can only spread and be used is a society where most people actually prefer the soulless and the mediocre; at least, when this is convenient.
On top of this; Western Civilization is led and inhabited by shoals of Godless materialistic people who care absolutely nothing for Truth, Beauty or Virtue; are indifferent to quality and even functionality.
These will implement AI even when it is functionally inferior or much worse, than existing systems and persons.
It is a measure of our civilizational and personal corruption that AI is being so rapidly and pervasively imposed upon The West - and why so many, at so many levels, have embraced it.
4 comments:
There was an interesting comment on this post (https://notionclubpapers.blogspot.com/2021/02/review-of-tolkiens-modern-reading-by.html) from Wurmbrand who read a book about Tolkien's reading (Cilli's Tolkien's library). He thought it was telling him something new, but its source was just an article he had already written.
What's disappointing is that even people who are knowledgeable and motivated seem to be embracing so-called AI. But at best all it can do is reflect back to people what they already know (if not making things up), but without even presenting its sources, so it can't even be checked like in that case of the Tolkien's library book.
@NLR - I do not myself find this surprising (although it is of course disappointing).
As you know from my 2012 book "Not Even trying" I regard science, and also academia, as having died-out at the professional (or large scale) level about a generation ago (i.e. c 30 years).
All that is left now (in a few areas) are a few amateur/ unofficial individuals, mostly working alone or with a handful of others.
One of the main attractions of this technology is perpetual novelty (novelty by recombination). You made a point last year (https://charltonteaching.blogspot.com/2024/07/whats-point-of-average-and-bad-work-in.html) about how consumption of bad works is about novelty and diversion. So-called AI is that to an even greater extreme.
@NLR - Yes. AI would only be of interest when aesthetic taste has been so degraded that almost every major media event is a sequel, crossover, pseudo-adaptation and the like. People who occupy "creative" positions in the hierarchy are not even trying to be genuinely creative nor to make genuine beauty. The opposite, indeed.
The same applies (mutatis mutandis) in the bureaucracies of science and academia wrt Truth. For the past few decades, the official and leadership negativity towards truth (e.g. disgust, mockery, even fear) has been palpable, and at times explicit.
Post a Comment