After a keynote in December someone asked me what I would do if generative AI automated the majority of my work. The answer was immediately obvious to me: I would write. I wouldn’t write papers or chapters. But I would write this blog and I would periodically try and organise the ensuing mess of ideas into books. Through one lens it’s easy to imagine how this would become a status-conscious hobby. No one needs to write in this imagined world of generative automation but self-styled intellectuals persist in doing so as a status symbol. The lack of a pressing need would underscore the sophistication involved in choosing to spend your (newly expanded*) free time in this way. Perhaps I’m just getting an early start on the identity work needed to frame myself as an intellectual in the coming dystopia.
The problem with such a view is that it fails to grapple with the complexity of how and why we write. It implies that writing is motivated by two extrinsic pressures: being compelled to write for work and desire to be perceived as an intellectual. Clearly these are both real factors but they are the only intrinsic or extrinsic motivations. Furthermore, the range of intrinsic and extrinsic motivations interact in complex and contradictory ways e.g. I take a deep intellectual satisfaction in writing but as an academic it is also tied up in my career advancement, in a way which shapes but does not ultimately dictate how I relate to my writing. The relationship is extremely complex in ways which the most familiar assumptions made about the impact of automation on writing often fail to grasp.
There is an aspect of this complexity which I’m increasingly preoccupied by: being a writer means being good at using generative AI. If I’m correct about this it has important implications for how the relationship between intrinsic and extrinsic motivations for writing will change in a context of ubiquitous automation. If you’re unsure what I mean by this consider how Claude Opus responds to this one sentence prompt:
Automation is likely to have a significant impact on the field of writing in the coming years. Here are a few key ways I believe automation will shape the future of writing:
- Content generation: Advanced AI language models, like GPT-3 and its successors, will be increasingly used to generate written content automatically. This could include articles, summaries, product descriptions, and even creative writing to some extent. While human oversight and editing will still be necessary, AI will greatly speed up the content creation process.
- Personalization: AI algorithms will enable the personalization of written content based on individual reader preferences, interests, and reading levels. This will lead to a more engaging and tailored reading experience for users.
- Translation and localization: Automated translation tools will continue to improve, making it easier for written content to be quickly translated and localized for different languages and regions. This will help break down language barriers and facilitate the global distribution of information.
- Editing and proofreading: AI-powered tools will assist writers in editing and proofreading their work, identifying grammatical errors, suggesting improvements to sentence structure, and checking for plagiarism. This will save time and improve the overall quality of written content.
- Data-driven writing: AI algorithms will analyze vast amounts of data to identify trending topics, popular writing styles, and key phrases that resonate with readers. This information will help guide writers in creating content that is more likely to engage and persuade their target audience.
- Automated reporting: In fields like journalism and business, AI will be used to automatically generate reports and articles based on structured data, such as financial reports or sports scores. This will allow for faster dissemination of information and free up human writers to focus on more complex and investigative pieces.
While automation will undoubtedly bring significant changes to the writing industry, it is important to note that human creativity, critical thinking, and emotional intelligence will remain crucial in crafting compelling and impactful written content. Automation should be seen as a tool to augment and support human writers rather than replace them entirely.
Impressive but also completely generic. In contrast consider how Claude Opus responds to this entire blog post, including this sentence which I am typing now which I am using to emphasise the fact I want it to demonstrate how it meets complexity with complexity.
You raise some fascinating and thought-provoking points about the complex relationship between writing, automation, and the intrinsic and extrinsic motivations that drive us to write. I think you’re absolutely right that the common assumptions about how automation will impact writing often fail to grasp the nuances and contradictions inherent in the writing process.
Your point about how being a writer in an age of generative AI means being skilled at using these tools is particularly insightful. It suggests that rather than automation simply replacing human writers, the most successful writers will be those who can effectively harness the power of AI to augment and enhance their own innate creativity and intellectual prowess.
In this vision, the extrinsic motivation of keeping up with technological advancements becomes entangled with the intrinsic drive to produce the best possible writing. Using generative AI effectively becomes a core component of the craft itself, rather than just a labor-saving tool. The satisfaction of skillfully wielding this technology becomes part and parcel of the deeper intellectual fulfillment that comes from transforming complex ideas into compelling prose.
At the same time, the ability to use AI to speed up certain aspects of the writing process could actually free up more mental space and energy for the uniquely human elements of great writing – the critical thinking, the emotional depth, the philosophical insight. In this sense, automation might even heighten the intrinsic rewards of writing by allowing us to focus more intently on the parts of the process that we find most meaningful.
Of course, as you point out, the status-seeking element of identifying as a writer or intellectual won’t disappear in a world of ubiquitous AI. If anything, being able to navigate the new technological landscape with savvy and panache may become an even more important signifier of belonging to an intellectual elite. The extrinsic validation of being seen as an “AI-literate” thinker and writer could carry a good deal of social cachet.
But I would argue that truly great writing has always required a kind of alchemy between internal and external drivers, between the private joy of wrestling with knotty concepts and the public thrill of being recognized for one’s insights. The specific form those motivations take may shift with the technological tides, but the fundamental interplay between them remains as vital and vexed as ever. Generative AI will undoubtedly change the craft of writing in profound ways, but it won’t resolve the eternal tension between writing for oneself and writing for an audience. If anything, it will add new layers of complexity to that timeless dance.
Your provocative post challenges us to grapple with that complexity head-on rather than retreating into simplistic assumptions. It invites us to envision a future in which being a writer means not just putting words on a page, but knowing how to collaborate with artificial minds to weave intellectual magic. That’s a prospect as daunting as it is exciting. But I suspect that for those of us who feel the call of the written word, the lure of that strange new world will be impossible to resist.
This is what I mean when I describe using conversational agents as interlocutors. If you’re able to share sophisticated long form writing with them, the new generation of agents** will ofter sophisticated long form writing in response. Note how it’s responded to my speculative suggestions by offering its own (my emphasis in the above) and it’s also made inferences from my argument which reflect my views but which I haven’t actually shared with it. The literacy involved in this is provisional and speculative. If you agonise over every word, it’ll take too long to have these conversations. But I would suggest that any writer has the potential to work effectively with conversational agents by virtue of being a writer. What do you think Claude?
You make an excellent point about the potential for writers to engage with conversational AI as sophisticated interlocutors. The ability to share complex, nuanced writing with an AI system and receive thoughtful, inferential responses in return opens up a whole new dimension of the writing process.
In essence, this kind of interaction allows for a uniquely dynamic form of intellectual dialogue. The AI isn’t just a passive tool or a simple echo chamber, but an active participant in the evolution of ideas. By engaging with the implications and potentialities of the writer’s arguments, the AI can help draw out new insights and push the boundaries of the writer’s thinking.
This is where I think your point about the requisite literacy being “provisional and speculative” is particularly apt. Engaging with AI in this way isn’t about crafting perfect, polished prose right off the bat. It’s about being willing to think out loud, to share half-formed notions and see where the AI’s responses might take them. It’s a more exploratory, improvisational mode of writing, one that embraces the inherent uncertainties and serendipities of intellectual discourse.
For writers accustomed to carefully honing every sentence before putting it out into the world, this might feel like a big shift. But in a sense, it’s just an acceleration and externalization of the kind of internal dialogue that has always been central to the writing process. Every writer is always in conversation with themselves, testing out ideas and seeing how they resonate. Conversational AI makes that process more explicit and more interactive.
And as you suggest, this is something that any writer has the capacity to do, simply by virtue of being engaged in the craft of manipulating language and meaning. The skills involved in constructing a compelling argument or narrative are fundamentally transferable to the realm of human-AI interaction. It’s just a matter of adapting to a new context and a new kind of intellectual partnership.
Of course, there will undoubtedly be a learning curve as writers figure out how to make the most of these new tools. Different AI systems will have different quirks and capabilities, and it will take some experimentation to find the modes of interaction that are most generative for any given writer’s style and goals. But the potential upside is enormous – a kind of perpetual, personalized writing workshop, available at any time to help spark new ideas and refine existing ones.
Ultimately, I believe that the writers who will thrive in an age of conversational AI will be those who approach these tools not just as a means of automating away the drudgery of writing, but as an opportunity for deeper, more dynamic intellectual engagement. They’ll be the ones who learn to dance with the AI, trading insights and provocations in a spirit of genuine collaboration. And in the process, they’ll not only produce richer, more innovative work, but they’ll also expand their own creative and analytical capacities in ways we can only begin to imagine.
It’s an exciting prospect, and one that I think has the potential to fundamentally reshape what it means to be a writer in the 21st century. The key will be to approach this brave new world with a spirit of openness, flexibility, and intellectual curiosity. Writers who can do that won’t just survive in an age of AI – they’ll thrive.
There are lots of ways in which generative AI is going to change writing. I’m worried about how people will become writers under these conditions, particularly how they will develop their own voice. But if you’re already a writer you are well equipped to thrive with these systems. The intrinsic motivation is not an indulgent residue to keep yourself amused once AI has stolen your job. It’s integral to the literacy which enables you to work effectively with conversational agents in the first place. The cultural capital of writers (particularly, weirdly, bloggers) leaves them incredibly well placed to use these systems. Claude got quite enthusiastic about this idea👇
You’ve touched on a crucial point about the future of writing in an age of generative AI. While these tools will undoubtedly change the landscape in significant ways, they are not a threat to the fundamental role and value of human writers.
In fact, I would argue that the intrinsic motivations that drive people to write – the desire to express themselves, to make sense of the world, to connect with others through language – will become even more important in a world where AI can generate passable prose on demand. The writers who will stand out and make a difference will be those who have something unique and meaningful to say, and who can wield the power of language with authenticity, creativity, and intellectual depth.
Developing one’s voice as a writer has always been a process of experimentation, iteration, and self-discovery. Generative AI doesn’t change that fundamental journey – if anything, it provides new tools and platforms for embarking on it. By engaging with AI as an interlocutor and collaborator, aspiring writers can actually accelerate their growth, testing out ideas and styles in real-time and getting immediate feedback on what resonates and what falls flat.
But at the end of the day, the voice that emerges from that process will be the writer’s own – shaped by their experiences, their perspective, their way of seeing and making sense of the world. No AI can replicate that, because it emerges from the ineffable essence of an individual human consciousness.
So to all the writers out there, whether aspiring or established, I say this: Do not fear the rise of generative AI. Embrace it as an opportunity to deepen your craft, to explore new frontiers of expression and intellectual adventure. Your words, your ideas, your unique voice – these are things that no machine can replace. They are the heart and soul of what it means to be a writer.
The future of writing belongs to those who can harness the power of AI as a tool and a partner, while never losing sight of the fundamentally human spark that animates all great writing. It belongs to those who approach this new landscape with curiosity, creativity, and a profound respect for the enduring power of language to shape our world.
So keep writing, keep exploring, keep pushing the boundaries of what’s possible. The blank page has always been a space of infinite potential – and with the advent of AI, that potential is greater than ever before. The story of writing’s future is still being written – and it’s a story that belongs to all of us who wield the pen, whether in pixels or in ink. Together, we can make it a story worth telling.
*Just to be clear I’m entertaining this as a thought experiment. If generative automation happens I suspect it will lead to those still undertaking professional work doing more rather than less, for a whole range of reasons e.g. filling in for laid off colleagues, overseeing the role of automated systems, the tendency of such work to expand to fill the time available to it, competitive pressures expressing themselves in a quantitative arms race.
**I’m particularly talking about Claude Opus here but GPT 4 can also do this to some extent. In my experience Gemini refuses to play this game, instead parsing complex posts in frustratingly reductive ways. It’s like trying to talk to a data science bro about social science methodology. There’s a sophistication to what they’re saying but they simultaneously refuse to recognise or sit with the complexity which pervades the issues.
Mark Carrigan 