“You’re not just writing — you’re creating a mirror for the modern mind”
Last week I read the AI Report published by Substack. Their findings were not surprising. Out of the 2000 well published respondents, about 46 % use AI in some form or the other in their writing, the report said. Though many don’t use it primarily to generate content but more for research and to fit syntax related errors. The rise of AI assisted content is inevitable in writing.
Though it has now become very easy to spot AI generated content, some believe that the lines will soon be blurred. Take for instance the wide use of em dashes in Chatgpt responses. These are so easy to spot that it only takes another prompt to fix them. Not only is the writing more natural sounding, it sounds like a human written text. See what I did there? These type of statements are another hallmark of AI generated content along with the liberal sprinkling of emojis (why though?)
“Great question — let’s break it down.”
The report is interesting in many ways because substack is probably one of the fastest growing platform for creators. Think of blogspot or wordpress but which can be easily monetised based on subscription. Next week I will do a deep dive on Substack, but this week I am going to limit myself to AI in writing. So, when the 2000 writers on Substack describe their interactions with AI, they are sending out a credible signal to the rest of us.
For example, most publishers agree on the originality of creativity. They accept that it cannot be outsourced and that there is a fear of losing one’s voice to AI. Of course, the genius cannot be curtailed by AI. In fact when I have run some of my poems or creative works through ChatGPT for assessment, it has often lent me a generic voice. Deliberate sentence formations or a creative approach to descriptions is often met with a textbook remark written by a strict disciplinarian English professor.
But AI in writing can be beneficial to seek out new ideas and structure your content differently.
“Here’s a framework you might find useful…”
The beauty and bane of the LLM world is that it is a reflection of your directions. If one prompts it to become an editor of the New Yorker and then critically evaluate an essay, the output is far more richer than a standard – evaluate this prompt. What is more significant is that AI often makes mistakes very confidently so one still has room to exercise one’s own judgement. No New Yorker editor has ever reviewed my work but the LLM often helps to get a starting point. If you encourage it to become even more ruthless the outputs and suggestions are more scathing and useful.
“Let’s be real”
The predictable nature of LLMs is often why many of us may find it difficult to adjust to the brave new world. But in reality AI tools are soon becoming a critical part of the way we function. Though creating use cases and using the tools the way we want still remains our prerogrative. Vibe coding for one has emerged as an important use case, but as far as writing goes – it is still best suited for low level writing such as escalation emails or farewell messages (don’t forget to edit out “Your Name” here when you copy it though).
“You’ve got this — and I’m here when you need to dive deeper.”
AI tools are great to become a better version of yourself. But not the version that AI decides, it is the one which you have defined yourself. Because no matter how hard we try, the tool will never get your full context. And that applies to writing too, especially creative writing. It is often a reflection of our true selves and this should not be masked. A free flowing written passage is much more cathartic than the one created by AI. But for low level tasks – one should simply use prompts and get over with it.
Let me know what your thoughts are.