In an Egyptian palace long ago, the god Theuth presented his groundbreaking invention – writing – to King Thamus. Theuth believed it would enhance memory and wisdom. Unexpectedly, the king rejected it. Writing, he argued, would weaken people’s memory. Plato captured the tension well: technological leaps empower us, but they also make once-essential skills obsolete. Writing – now fundamental to civilisation – was once seen as a threat to human thought. History is littered with such fears, from sailors steering their ships by the stars to engineers tapping away at their slide rules. Today, it is artificial intelligence (AI) that’s reshaping how we communicate. And the question looms: Is the written word next in line?
Homo sapiens has been around for roughly 300,000 years. For most of that time, we got by without writing, learning instead through imitation, apprenticeship and storytelling. Oral traditions preserved astonishingly complex knowledge: Australian Aboriginal songlines mapped out waterholes in the desert, while Mali’s griots could recite centuries of dynastic history. Half a century ago, anthropologists Jack Goody and Ian Watt emphasised how these oral traditions structured social memory in preliterate societies, providing stability and coherence over generations.1 Spoken, not written, language was the bedrock of knowledge transmission.
Then, about 5,000 years ago, we began to write – and that changed everything. Writing allowed us to externalise our memory, paving the way for lengthy arguments, bureaucracy, science and literature. Over time, literacy became synonymous with intelligence. Yet, like every major technological shift, writing came at a price: it weakened our natural memory, reshaped cognition, and favoured slower, more linear thinking. The economic historian Joel Mokyr describes such shifts as ‘technological anxiety’ – society’s recurring fear that innovation might degrade essential human skills.2 Today, we find ourselves at another crossroads. The act of writing – as we know it – may be nearing its expiry date.
AI is already transforming how we communicate. Voice assistants like Siri and Alexa let us talk to machines, but they’re still too clunky, too slow for natural back-and-forth. Even the most advanced models struggle with high-stakes conversations. A captain on a sinking ship does not have time for an AI to assess urgency before acting – crisis decisions demand instinct, improvisation and a fine-tuned sense of human emotion. These subtle cognitive processes are forms of ‘tacit knowledge’ – nuanced human expertise notoriously difficult to codify. For now, that is still out of AI’s league.
But the tech is improving at lightning speed. As discussed on The Cognitive Revolution podcast, new models like Hume AI’s Octave and Google’s NotebookLM are faster, and increasingly capable of conveying emotion. AI is already negotiating freight contracts, fielding customer service calls and even offering executive coaching. Persuasion, storytelling and improvisation may still be human strongholds – for now – but the gap is closing. Recent breakthroughs in ‘transformer models’ – the computational backbone of tools like ChatGPT3 – allow AI to sound less mechanical.
It is already shaping the way we work and play. According to Edison Research’s Infinite Dial data, podcast listenership has surged since 2013. South Africa joined the survey later, but as the chart shows, the share of podcast listeners rose from 6% in 2019 to 20% in 2021 – a staggering jump in just two years. The same trend is unfolding in business and industry, where AI-driven voice interfaces are transforming customer service, contract negotiation, and even creative work. AI already helps with brainstorming and content creation – not just text, but dynamic, interactive speech.
Walter Ong coined the term ‘secondary orality’ to describe the revival of spoken communication in the digital age. Podcasts, voice notes and video calls are increasingly replacing text messages. This shift reintroduces elements of performance, spontaneity and collective participation that were central to ancient oral storytelling traditions. But AI is taking things one step further: it is building a world where writing is no longer a necessary bottleneck to thinking.
Technological shifts have always displaced certain skills. Gutenberg’s printing press rendered scribes obsolete. Calculators made mental arithmetic redundant. GPS weakened our sense of spatial navigation. In the same way, writing may soon join the list. AI can already write essays, draft reports and compose emails. But the true revolution will arrive when we no longer need pen or keyboard to shape, refine or communicate ideas – when the spoken word, mediated by AI, takes its rightful place once more.
Importantly, AI’s growing multilingualism means this shift won’t be confined to English. Until recently, most speech models were trained in a handful of dominant languages. But new research in the Journal of Machine Learning Research shows that AI can now generate natural speech in languages as varied as Mandarin and Swahili.4 The implications are profound. This technological multilingualism, argues sociolinguist Alastair Pennycook, may accelerate linguistic diversity rather than diminish it.5 AI-powered rhetoric means that persuasive language and high-level dialogue may no longer be constrained by language barriers. Spoken conversation could become the dominant form of global communication.
As AI takes over everyday writing tasks, rhetoric – the art of persuasion – is set for a renaissance. Long before writing, philosophers and leaders shaped societies with their arguments. The Ancient Greeks saw rhetoric, alongside logic and ethics, as the foundation of education. The future will demand the same skills: the ability to persuade, to captivate, to use speech not merely as a tool, but as an art form.
The communicators of the future will be skilled orators, weaving together techniques from storytelling, digital media and AI-mediated rhetoric. This shift may renew emphasis on the performative and ritualistic aspects of human interaction. The goal remains the same – effective communication – but the medium is shifting. Writing was transformative, yes. But AI’s conversational capacity may offer something richer. It promises immediacy, interaction and a return to the cognitive world our ancestors inhabited.
AI might just make us more like them. Perhaps soon, we’ll gather around a digital fire – talking, debating, learning from each other. Not through static text, but through conversation and rhetoric. The pen may yield to the tongue (or the microphone), but knowledge will continue to grow.
Welcome (back) to the Age of Orality!
This is an edited and translated version of my monthly column, Agterstories, on Litnet. To support more writing like this, consider subscribing for a paid membership. The images were created using Midjourney.
Goody, J. and Watt, I., 1963. The consequences of literacy. Comparative studies in society and history, 5(3), pp.304-345.
Joel Mokyr. The Gifts of Athena: Historical Origins of the Knowledge Economy. Princeton, NJ, and Oxford: Princeton University Press, 2002
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł. and Polosukhin, I., 2017. Attention is all you need. Advances in neural information processing systems, 30.
Pratap, V., Tjandra, A., Shi, B., Tomasello, P., Babu, A., Kundu, S., Elkahky, A., Ni, Z., Vyas, A., Fazel-Zarandi, M. and Baevski, A., 2024. Scaling speech technology to 1,000+ languages. Journal of Machine Learning Research, 25(97), pp.1-52.
Alastair Pennycook. Global Englishes and transcultural flows. Routledge, 2007.
I think we would do well to give this dedicated consideration. From my very limited, mostly academic, perspective, it seems that there are emerging trends to suggest that the "Age of Orality" is very real. More and more, I see invitations to online discussions, webinars, etc. Some, in my experience, are revaluing discussions as ways of data creation and collection. This demand for human interaction is exciting. Indeed, an interesting read - well done!