In Ancient Greece, stonemasons used to etch inscriptions using the boustrophedon writing system. This is a style of presenting text that zigzags; left-to-right on one line, right-to-left on the next. It literally means “as the ox turns”, just like an ox ploughing a field, pivoting at the end of each furrow to begin another in the opposite direction.

This didn’t last. Just as the ox was slow and laborious in its turns, so too was the cognitive strain placed on readers as they tried to access the text. You get tired. Try reading this quickly:

THE OX IS PLOUGHING THE FIELD. IT WILL FINISH BY DUSK

KSUD YB HSINIF LLIW TI .DLIEF EHT GNIHGUOLP SI XO EHT

It hurts my brain.

Our brains are wired for rhythm and consistency, not for abrupt shifts. Over time, we developed simple, more predictive systems – left-to-right in Latin, right-to-left in Arabic, vertical in Traditional Chinese or old Mongolian scripts.

Essentially, this story reveals a simple truth: writing and reading have always been shaped by our own biological constraints, cultural solutions and to fit the technology of the time. But now, technology is taking reading in a whole new direction.

Our Biology: Wired for Rhythm and Sequence

Our brains process written information incrementally. Our eyes only take in a few words at a time. We prefer linearity, whether left-to-right, right-to-left or up-to-down. This provides rhythm and predictability. The boustrophedon’s back-and-forth motion did not work because it disrupted this pattern. It overloaded working memory and forced an awkward reorientation.

The way we process information shaped the evolution of reading systems. We expect

rhythm, sequence and consistent directionality. Just as the ox ploughs one straight

furrow after another, so do we when we read: line by line and page by page.

The Culture Layer

Culture imposes a layer of meaning on biology. While the latter anchors how we process

text, culture refines the structure. Writing systems through the ages arose from practical

and social contexts:

·       Cuneiform evolved to suit clay tablets and the stylus

·       Vertical Chinese script flowed naturally on scrolls

·       The codex and then the book encouraged horizontal, left-to-right reading.

Beyond scripts, culture also shaped reading habits:

·       Western readers scan for headings, bullet points and direct calls to action

·       Other cultures, e.g. high-context societies, favour narrative explanations, formal tone and layouts that emphasise balance and subtlety.

Similarly, how we navigate texts is also culturally bound. Western readers begin in the top-left corner, while readers in right-to-left cultures look to the top-right. Ways of reading reflect both universal biology and local cultural values.

AI: The Reader Without Limits

Enter AI – a reader that is not tethered to biology. AI does not read the way that humans do. It dissects paragraphs into ‘tokens’ and processes them in parallel. It uses mechanisms to spot patterns and relationships that we humans miss entirely.

While we graduate from word to word and sentence by sentence, AI bounds across the textual landscape, making connections between distant ideas – joining the dots – instantly: a far cry from the ox’s slow, deliberate ploughing.

Though we state that AI is untethered to human biology – any biology – it is nonetheless rooted in human culture. AI models are all trained on text that are steeped in our biological and cultural preferences and biases. I understand why writers like Yuval Noah Harari in his book Nexus prefers the term  ‘Alien Intelligence’ to ‘Artificial Intelligence’ – because it is alien to our decision-making processes. If so, it is alien intelligence trained on very human soil.

The Implications of AI on Cultural Expectations of Processing Texts

The divergence between AI and humans creates a tension, when it comes to processing information.

·       If AI can read and digest information faster and better than us: texts, reports, whole datasets and generating text, can it still preserve the nuance and cultural expectation embedded in human-authored texts? Would it go beyond what a human can do and integrate cultural expectations before the text is even developed?

·       Alternatively, AI-produced content leans towards efficiency and favours clean, universal layouts, but it is obviously based on a Western model of information design, structure, tone and flow. Therefore, could it forever alienate readers who have specific cultural expectations when processing texts?

·       Would this lead to cultural flattening? As AI mediates more communication globally, will reading conventions become homogenised? Would cultural diversity in reading habits still persist?

A Thought Experiment: What If We Outsource Reading?

It might come as a surprise that we have already outsourced some of our short-term and working memory functions to digital tools. This is known as the Google Effect. This is the tendency to forget information that we can easily look up online. It has had a direct effect on our collective short-term memories over the past twenty years.

What if the same thing happens with reading itself?

Imagine a world where AI reads for us: summarising, filtering, prioritising information. AI becomes the one who reads for us, decides what we have to read and gives us summaries according to what an algorithm decides what is important for us. What would happen to human cognition?

·       There could be a vast reduction in deep reading, intellectual curiosity, serendipitous discovery. If AI is doing all the heavy lifting, we could lose the patience and cognitive stamina for deep, critical reading.

·       An erosion in critical thinking. Much of critical thought comes to us through the slow process of reading texts that challenge us and then spending time thinking about what we have read. Yet, if AI delivers ready-made insights, would the muscles of analysis, doubt and scepticism atrophy?

·       If we depend on machine filters, where AI decides what we see, could human understanding narrow, shaped by algorithms that we don’t fully understand (or at all)?

·       If AI provides audio summaries of what it reads for us, would we need to learn to read at all?

We could move from a post-textual to a post-literate world, where audio and video completely replace the need for most people to read. But somebody would need to construct the messages in the first place. Would literacy then be the domain of the 1%, just like with the scribes in Ancient Egypt? The literate becomes the priesthood, telling us what and how to think. It is unthinkable that the removal of literacy could be part of our quantum stride forwards. It feels like a paradox. It feels wrong.

Thus, the ox might stop ploughing altogether, while the machine not only races ahead but decides which fields matter and which do not.

But for now, the fields remain ours to plough – in straight lines, one furrow at a time. The machine, however, is ripping out the hedgerows that define our boundaries. Profound change could come swiftly and soon.

Thank you for reading,

Mark Gibson

Leeds, United Kingdom, March 2025