The retained blue GPTOnline cover is an editorial graphic for these five writing tips, not a ChatGPT interface screenshot or evidence of expert testing.
Darlings, it’s time to transform ChatGPT into your personal writing genie!
I knooow you’ve been feverishly punching prompt after prompt into that box, expecting Shakespearean prose to fill your screen. And yet…nada! Just some robotic ramblings not fit for human eyes.
Well never fear my dears, “Auntie Adriana” is the playful editorial voice for these five writing tips, not a claim about a real expert’s identity or firsthand product testing. Grab a drink and settle in. Class is in session!
Rule #1 for prompting writing brilliance from ChatGPT?
I know your brilliant mind is overflowing with eleventy-billion ideas you’re positively dying to see in writing. But hold those beautiful horses! If you unleash a tangled web of complex prompts all at once, poor ChatGPT gets more confused than my 102 year old granny trying to work her iPhone.
ChatGPT can respond to multi-part requests, darlings, but breaking a complicated writing job into smaller steps can make errors easier to spot. Give ChatGPT one single writing task to focus its neurons on before promptly moving on to the next micro-prompt.
I like to pretend I’m a manager giving crystal clear direction to my new intern ahem ChatGPT. “First I need an opening hook grabbing readers’ attention. Make it punchy!” Send that off, then review the output and provide feedback before approving the hook and giving the next teeny tiny assignment.
It’s tedious I knooow, but think about it as giving toddler-sized portions to a child just learning how to handle grown up food! Tiny bites prevent big tummy aches (or rather, crappy writing). This incremental prompting approach lets you nip any errors in the bud early on, instead of trying to unravel a giant knotty mess after the fact!
Which brings me to rule #2 Cut the Ambiguity & Give Concise Instructions!
If you’re still not satisfied with what ChatGPT is serving up sentence-wise, chances are your prompts are more vague and nebulous than a palm reader forecasting my love life!
Look honey, no judgment here…but vague questions can lead ChatGPT to produce answers that miss your intent. Simply not gonna cut it if you’re looking for prose as lustrous as Rapunzel’s hair!
To give the draft a clearer direction, you need prompts cut straight from crystal. Define all the terms, set tight expectations, and force ChatGPT to hyper-focus on precisely what elements you expect in its response, down to the last letter.
State the constraints that matter, and leave creative choices open where you want alternatives. Make sure ChatGPT understands terms the same way you do. Define genre, audience, formatting – the works!
You can also ask ChatGPT for feedback on a draft prompt’s clarity and missing context. Think of it as an assistant for your assistant, not independent verification or a promise of free access to every feature. Meta, right?
Now for Rule #3…Set the Context, Hunty!
Context is queen when prompting writing greatness from ChatGPT, henny! As in, lavish that AI with more background details than a Real Housewife in a confession booth!
Plop this digital assistant right in the thick of a fleshed out scene, rich with sensory details and crystal clear objectives. Pretend you’re screenwriting the next Hollywood blockbuster smash, with ChatGPT playing the sassy yet loveable protagonist.
Give ALL the supporting deets – explain relevant backstories, lay out comprehensive premises, unpack motives…the works! Relevant context can reduce ambiguity, but it does not eliminate errors or omissions. Re-read and triple check that your AI BFF truly understands its mission before letting those neural networks fly solo!
Which segues smashing into Rule #4: Divide the Labor Between Human & Robot!
Alright, so you’ve laid down the landscape like the world’s greatest set decorator prepping for film day #1. Now it’s time to shout “Action!” and watch the magic unfold!
…Well, your part of the magic, anyway. This is when you decide whether you personally want to supply the creative ideas while leaving ChatGPT to help draft or organize research notes…OR, if you rather prefer providing meticulous editing while ChatGPT supplies the raw prose for you to refine.
Be crystal clear in divvying up assignments between human and AI based on each of your specialized skillsets. Did I say crystal clear? Dahhling, I meant freaking DIAMOND clear!
Explicitly outline what corner of creativity belongs to you, and which drafting tasks you want ChatGPT to handle. Verify factual claims and citations yourself before publishing. Confusion breeds chaos in this partnership, so banish it from the kingdom!
Now for my personal fave: Rule #5 – Become a Feedback Fiend!
Go ahead and banish any fairytale fantasies of ChatGPT delivering Pulitzer-prize prose on the first try. We ALL have room for improvement, AI honey bunch included!
Instead, put on that unforgiving editor hat and scrutinize each draft with the determination of an Olympic judge. Figure out exactly which phrases sent your writing sideways, then servant-call ChatGPT for a do-over with highly specific corrections.
Love this opening hook but next paragraph needs to foreshadow Annie’s inner turmoil more subtly without giving everything away.
Get granular, get meticulous, get micro with your feedback! Wash rinse repeat this process, prompt by prompt, until the draft meets your brief and passes your own review. Consider it a cheeky writers’ tango pushing creativity to uncharted heights!
Context: These exercises come from a December 2023 article, not a measured productivity or business-results study. OpenAI’s current prompting guidance supports clear instructions, relevant context and iterative review; its accuracy guidance explains why outputs still need checking.
