There’s recently been news about a developer who has managed to “build free, open-source alternatives to Adobe’s creative apps” using Claude, Anthropic’s AI assistance model. Regardless of their current effectiveness, which is being debated, this development has sparked some discussions about how AI can effectively surpass ridiculously over-priced corporate paywalls on widely-used creative tools, by providing public alternatives.
These conversations are clearly connected to wider ideas about how AI and its widespread dissemination can help democratise access to knowledge, leading to some remarkable intellectual and creative progress across the globe, or at least in the US-European region, as it so often seems to be not-so-subtly implied.
Unfortunately, this frequently very shallow and poorly examined ‘innovative potential’ is distracting us from its inherent flaws, and other cheaper, and generally less harmful, options to improve knowledge-accessibility.
It is first important to recognise that providing a fair and easy access to information is not, and has never been, the purpose of these AI tools. It is solely to make AI corporations, and their owners, a bunch of money by pushing egotistically ballooned narratives about the excellence of these technologies, to intrigue investors, and create an illusion of this business proposal’s essentially predetermined success. They are displayed as flawless and clean and perfect, intangible machines, symbolising humanity’s innovative future, because that all sounds very pretty when you don’t think about it too much, and pretty, uncomplicated things are easily marketable, without actually needing to provide a social good.
The solution open-source AI provides, as promoted by its proponents, reminds me a lot of a section in Zamyatin’s 1920 novel We, about the “musicometer,” which allows anyone to “produce up to three sonatas an hour” by “simply turning the handle”— an alternative to seeking and working one’s “inspiration,” which makes creating music very difficult by comparison. I admit, however, that this isn’t an ideal parallel; AI is also an artometer, and ideaometer, and thoughtometer. The point is, in the end, this dystopian solution only takes away from human possibilities by providing a passive (and so insincere) mode of knowledge-engagement, while stealing attention from more hopeful alternatives that could actively involve people in their efforts to seek knowledge.
The clearest — but obviously far too dull for people to take seriously — example is increasing the number of well-funded, public libraries. Libraries can provide physical spaces for people to connect and socialise, with plentiful information, and free access. Next to the “staggering costs” of building an AI data centre and its incomprehensibly destructive environmental impacts, while providing little-to-no long-term jobs, I’d like to see an AI-fanatic embarrass themselves arguing against my suggestion.
Continuing with the suitable We comparison, it seems to me that an AI-centred method of democratising knowledge is more focused on having knowledge become “conquered i.e. organised and mathematised,” rather than truly ‘opened’ to the public. This is most notable in how it tends to steal from artists and academics without crediting them, or how it invents references out of thin air.
I feel that I’ve repeated the basic concerns about AI countless times at this point, be it in speaking with others or writing critical pieces on the subject, but I’d like to reiterate that AI tools are not objective, all-knowing beings with real intelligence. They are limited by what they can access. Some genius might counter: “Well, then, let’s just feed it all of the information we have!” Unfortunately, as much as its creators want you to believe it, AI is also not flawless, and physical alternatives are much less susceptible to having their content unknowingly shifted in an intangible data-space.
We can agree that some type of solution to the unchecked social and economic power of corporations is urgently needed, while critiquing this corporate-made tool that we believe to be seeing some emancipatory potential in. More tangible, people-oriented proposals will certainly provide better, long-term, and transformative improvements in democratising knowledge.
Photo by Nahrizul Kadri on Unsplash.
Related
Open-source AI is the wrong way to democratise knowledge
There’s recently been news about a developer who has managed to “build free, open-source alternatives to Adobe’s creative apps” using Claude, Anthropic’s AI assistance model. Regardless of their current effectiveness, which is being debated, this development has sparked some discussions about how AI can effectively surpass ridiculously over-priced corporate paywalls on widely-used creative tools, by providing public alternatives.
These conversations are clearly connected to wider ideas about how AI and its widespread dissemination can help democratise access to knowledge, leading to some remarkable intellectual and creative progress across the globe, or at least in the US-European region, as it so often seems to be not-so-subtly implied.
Unfortunately, this frequently very shallow and poorly examined ‘innovative potential’ is distracting us from its inherent flaws, and other cheaper, and generally less harmful, options to improve knowledge-accessibility.
It is first important to recognise that providing a fair and easy access to information is not, and has never been, the purpose of these AI tools. It is solely to make AI corporations, and their owners, a bunch of money by pushing egotistically ballooned narratives about the excellence of these technologies, to intrigue investors, and create an illusion of this business proposal’s essentially predetermined success. They are displayed as flawless and clean and perfect, intangible machines, symbolising humanity’s innovative future, because that all sounds very pretty when you don’t think about it too much, and pretty, uncomplicated things are easily marketable, without actually needing to provide a social good.
The solution open-source AI provides, as promoted by its proponents, reminds me a lot of a section in Zamyatin’s 1920 novel We, about the “musicometer,” which allows anyone to “produce up to three sonatas an hour” by “simply turning the handle”— an alternative to seeking and working one’s “inspiration,” which makes creating music very difficult by comparison. I admit, however, that this isn’t an ideal parallel; AI is also an artometer, and ideaometer, and thoughtometer. The point is, in the end, this dystopian solution only takes away from human possibilities by providing a passive (and so insincere) mode of knowledge-engagement, while stealing attention from more hopeful alternatives that could actively involve people in their efforts to seek knowledge.
The clearest — but obviously far too dull for people to take seriously — example is increasing the number of well-funded, public libraries. Libraries can provide physical spaces for people to connect and socialise, with plentiful information, and free access. Next to the “staggering costs” of building an AI data centre and its incomprehensibly destructive environmental impacts, while providing little-to-no long-term jobs, I’d like to see an AI-fanatic embarrass themselves arguing against my suggestion.
Continuing with the suitable We comparison, it seems to me that an AI-centred method of democratising knowledge is more focused on having knowledge become “conquered i.e. organised and mathematised,” rather than truly ‘opened’ to the public. This is most notable in how it tends to steal from artists and academics without crediting them, or how it invents references out of thin air.
How can Reform be pro-family and anti-childcare?
I feel that I’ve repeated the basic concerns about AI countless times at this point, be it in speaking with others or writing critical pieces on the subject, but I’d like to reiterate that AI tools are not objective, all-knowing beings with real intelligence. They are limited by what they can access. Some genius might counter: “Well, then, let’s just feed it all of the information we have!” Unfortunately, as much as its creators want you to believe it, AI is also not flawless, and physical alternatives are much less susceptible to having their content unknowingly shifted in an intangible data-space.
We can agree that some type of solution to the unchecked social and economic power of corporations is urgently needed, while critiquing this corporate-made tool that we believe to be seeing some emancipatory potential in. More tangible, people-oriented proposals will certainly provide better, long-term, and transformative improvements in democratising knowledge.
Photo by Nahrizul Kadri on Unsplash.
Share this:
Like this:
Related