I’ve been a full-time freelance writer for 21 years. Over the last 18 months, I’ve used ChatGPT to reduce my research time from several hours to seconds, help me draft, process, and prioritize dense information from the sometimes pages worth of detailed scripts I feed it, and even use it to test the most clickable headlines. I’m convinced it’s the most powerful “machine” of my lifetime so far.

That said, I never let AI write my emails, full articles, or blogs. I never let it rob me of my voice. I limit how much it reduces my ability to think while letting it do so much heavy lifting.

Why? Because I can’t make personal connections with a “mechanical middle-man” that doesn’t even understand language. In terms of intelligence, AI is moronic at best, an idiot at worst. It’s not even adolescent in its understanding of humans. It’s just the most amazing “fake it till you make it” machine ever created by brilliant humans.

Parallel computing is the real power of AI. Instead of doing one thing at a time like we do (i.e. serial computing), it uses millions of “helpers” or processors to break down a giant problem or question at once and then assembles in seconds what it predicts to be the best answer to satisfy humans, based on the information we gave it (or have made publicly available online).

It doesn’t really understand your relationship with the person on the other end of the email, article, or whatever else you’ve written. Maybe someday it will. Maybe one day it will become truly intelligent and better, more efficient, cheaper, and fully capable of expressing, writing, and dictating human emotions.

But until then, please use it what it’s best for: a massive spellcheck, predictive text creator, research assistant, and sometimes rough-draft generator on steroids.