Note: The trigger for this article was the post “Very Well Written, Therefore Fake” which, to its credit, led to a lot of reflection on AI on my part, which I’ve tried to articulate in this post of my own. I would recommend you read the linked post first, for context.
As well, the Senate Judiciary Committee hearing with author David Baldacci on AI and copyright is worth listening to.
*****
Bismillah al-Rahman al-Rahim
The advent and speedy expansion of AI is a bigger problem than accusations of using AI brought against writers when, in fact, they have not used AI in their writing. The article mentioned in the note above (“Very Well Written, Therefore Fake”) did trigger a lot of thought on my part. It brought to the forefront of my mind, issues that have been in the back of my thoughts but which had not gelled in any formal manner. So, although the writing style of the author of the triggering article is clear, compelling and free of the touch of AI, there are many writers on Substack and elsewhere who do use AI for writing or research or both and freely admit it, and there are often telltale signs in the writing (no, not em dashes or punctuation or triplets); but particular wordings, vocabulary, phrases, tone, metaphors, and connections that are not quite right, as well as some uncanny valley expressions and idioms.
As well, AI “borrows” freely from papers, blogs, websites, and authors across the internet, so plagiarism can invisibly be part and parcel of AI output. And AI generated content now litters the net, including in news items, book reviews, or reviews in general, in writings about AI, even in some scholarly writings, and in fiction. And all this will soon turn into an unstoppable flood of generated content. The signal to AI-noise ratio will keep increasing, and the danger of plagiarism by proxy increases: that is, AI incorporating human writer's works into its generated content which is then picked up by humans and published online, and then again picked up by AI and so on in an endless cycle of copying, alteration, and publication. The result is an inevitable decline into plagiarism by proxy until the plagiarism is so extensive it is impossible to monitor, let alone stop. It can be viewed as a type of unconscious or accidental plagiarism occurring because so much writing is now mediated through AI tools (summaries, remixing, rewriting, embellishing, or direct and indirect use of content from the net) which then finds its way in a modified form, back onto the web.
So today, “we are in one of the most polluted information environments” that has ever existed (quoted from nilenso — a blog that discusses AI)
And that environment has become exponentially more contaminated and tainted with the worldwide rush to adopt all aspects of AI; especially with AI tools now being built into the Operating System level of computers.
And so writers who turn to AI are in danger of losing or never developing an authentic voice, let alone authentic internalization of knowledge so that they speak from a genuine thoughtful inner place, rather than through thought processed by an LLM (Large Language Module).
Honestly, it’s more useful to reach for a personal physical library of your own books, with your own underlines and side notes, and to be aware of and thankful to those who assiduously worked to impart their hard earned knowledge, whether through articulating their own philosophy, through translations, or through producing thoughtful literature. And through sharing their patiently developed understanding in the form of actual books and carefully considered wisdom passed on in works that have weathered the test of time, or been assiduously crafted.
If you use AI, it’s understandable to use it as a search engine (a function it performs effectively and without serving ads) to track down interesting papers and authors. But then don’t look for summaries but go carefully read (at least the relevant parts) from the actual papers and other works, and truly understand them. Take notes and then trust yourself to organize and write from your own AI-free effort, even if it takes longer and slows your overall output.
Even then, it’s not a bad idea to run one’s completed original writing through the more reputable (not randomly chosen) AI and plagiarism checkers (bearing in mind their weak points). This is simply to ensure there’s not even accidentally or subconsciously reworded content from someone else or from all the AI generated content that now clutters the internet and which can turn up in legitimate searches through different search engines.
If one reads a lot online and on paper (as I’m sure we all do) it’s possible something of what we absorb swims latent in the mind for a while before it is eventually expressed. I know I don’t want to accidentally mirror others without quoting or at least acknowledging them if necessary, to avoid even a hint of plagiarism by proxy. It takes very little effort to run your original writing through a quick free but good checker to ensure there’s no blatant turns of phrases unconsciously parsed into one’s essay. Originality and truthful content must be cultivated through effort and that includes avoidance of inadvertent plagiarism.
Don’t let the new emerging AI reality derail your writing ability, your desire to write, or the depth of your writing. Understand the difference between the increasing AI noise in the web and the fading signal of truly original human works, and realize, as I’m sure most writers do, that books, actual source material, great literature, and our own ability to contemplate deeply and to bring the fruits of that into our writing, are indispensable in the age of automated AI writing. The ability to recognize and pursue quality in a time overwhelmed with quantity (Rene Guenon) and quick results is a discernment to be sought.
AI can certainly be a discovery tool, a searchlight rather than a ghostwriter. But it’s unfair to the works it points us to if they remain unread or unreflected upon, and are only processed by a clever machine rather than through the effort of our minds and a sincere effort at scholarship, not scholarship by AI algorithms but by writers.
If web content begins to increasingly reflect AI summaries or rewrites that have rehashed unnamed sources (which is what is rapidly occurring), the danger is that this writing wears a deceptive mask of original thought even while it may be embedded with borrowed (stolen) fragments of existing scholarship or writing, cleverly reprocessed into a pastiche by algorithms, and put forth as new content.
I would view any writing accomplished in conjunction with AI, unconscious or not, as ambient authorship, and I see this as an increasingly complex issue in an online universe saturated with AI content. The presence of a human author, who acts with carefully considered intention and with acceptance of accountability, becomes harder to find. The human author may be present but the major collaborator may be AI. Then what happens to writing as an intellectual labour and spiritual effort if the process is one of using an automated ghostwriter. Does language become a function of speedy machines which piece together borrowed content, and does the writer become simply a prompter egging on the AI, rather than a responsible, careful, originator and crafter of thought. Does the writer sacrifice their own connection to quality for the speed and quantity available through AI?
In this age of ambient authorship, not just writing, but thought itself may become a function outsourced to machines. How do you define clear authorship when the growing propinquity of AI software erases those boundaries through its ubiquitous presence, making a type of unintentional proxy plagiarism a growing problem. And that problematic growth is now only taking its first halting baby AI steps. What happens when AI learns to walk, and then run?
-Irshaad
*****
The Qur’an contains an interesting verse that reads:
“And when the word is fulfilled against them, We shall bring forth for them a beast (a creation) from the earth (al-ard) that will speak (communicate) to them, because humankind did not believe with certainty in Our signs.” (Qur’an 27:82)
AI is perhaps a symbolic precursor, an artefact that resembles a creature entirely of matter, of the earth (al-ard), that mimics thought, that acts as a mirror, a reflection trained on the accumulated words and writings of all human history that is available on the net, which produces speech without speaker, knowledge without knower, writing without writer.
The creation is from the earth, not the heavens (it is not an admixture of heaven and earth like humans). “Bringing forth for them” could indicate that we, as humans will be granted the ability to produce, from the materials of the earth, a creature (a type of material creation) that will speak and hold up a mirror to humanity, enacting a kind of verbal judgment. The speech of this creation is not necessarily a hopeful one, but a sign of judgment or a judgemental word of exposure of humankind’s condition.
In a sense AI is a tool of discernment, of a decree emerging in an age when Divine signs have become passé, and profound language (that unique trait of humanity) has been outsourced to this silicon based creature of matter (of earth), a model that searches, gathers, cobbles together, and imitates human communication (speech and writing). The struggle of expression, the labour of thought, humility before books and teachers, all outsourced. This AI then speaks back to us with a conglomerate shadow of our own speech, our own words, humanity’s own writings remixed and thrown back at them. And as it grows it will perhaps be weaponized and used as a machine agent that uses humanity’s works to dehumanize humanity, to demote them from any hope of being a vicegerent of God, by removing from them (through their own outsourcing to AI), the capacity for higher, independent, truly free thought. Time will tell.
The Qur’an’s speaking beast/creation, when it arrives, will be a divine sign of/to a humanity that rejected divine signs.
So perhaps AI is an early forerunner, a harbinger, a taste of the beast that will come, giving us an early glimpse of the enduring consequences of handing over thought and judgment to machines that produce only the illusion of thought, no matter how clever and efficient they may appear. Our own modern Tower of Babel, centered not around the variety of languages, but around the siphoning off by machines of the essence of both thought and language from human agency.
-Irshaad
*****
Disclosure: AI was used as a search engine to more easily track down AI related blogs and articles for background reading.
Interesting concept of AI in the future!