10 Dos & Don'ts of Writing with ChatGPT
To err is human,
To imitate humans is AI.
Think of the chatbots as improv actors taking a suggestion from the audience. They’ve already studied every SNL sketch — and now springboard new scenes off their vast repertoire + your direction. But just as an improv actor might tell a cringe joke, an AI might generate a response that doesn’t quite hit the mark. This is because improvisation, whether by a human performer or an AI chatbot, is inherently unpredictable and can lead to moments of brilliance as well as moments that fall flat.
Here’s another example for beauty bloggers and skinfluencers. Because AI reads everything on the internet, it can give you helpful medical facts from American Academy of Dermatology; however, it cannot exclude harmful beauty standards and hate speech, from Reddit for example. Answers from ChatGPT can be very biased and sometimes offensive. Always read before you publish.
AI is here to stay and help make our lives less soul crushing with daily tasks, not enslave humanity. That said, there are a lot of moments where it can go awry, from delivering misinformation that spreads on the web like wildfire to giving you creepy advice based in misogyny from toxic corners of the internet. Ensure you’re using the proper guardrails, fact checking, and editing every response.
What it can do
Research
Brainstorm topics
Write a blog post, social caption, subject line, or presentation
Answer questions
Interpret data
Read everything ever published
Give you a compliment
Reference the entire internet
Review a contract
Teach you new skills
What it can’t do
Fact check
Create new ideas
Write a promo email to save its life
Read your mind (so be specific)
Keep data private (so be careful)
Compensate writers and authors
Take you on a date…yet
Remove bias and hate speech
Provide sources
Hack your ex’s IG