I’ve seen it happen again and again: a writer, be they at a burgeoning news site or in an SEO content mine, is getting hit for sounding like AI. The tools that sniff them out, like Quillbot or ZeroGPT, purport to be able to tell AI writing from human writing with some degree of accuracy.
The results can be devastating, leading to lost jobs and frustrated writers. All because we trust AI detectors to work properly.
Trust us: they don’t.
AI writing is human writing. At least, it comes from human writing, and it is basically autocomplete with some degree of flexibility that masquerades as intelligence. Therefore, writing that comes out of an AI cannot be detected in any real sense.
The Media Copilot is reader- and affiliate-supported. If you click on one of these links, we might get a little cash if you like the product.
Incogni is a personal data removal service that scrubs your personal information from the web. Get 55% off with the code COPILOT.
Frase.io is a purpose-built AI writer that crafts SEO-focused articles, with detailed guidance on keywords and how to rank higher than competitors
Surfshark is a budget-friendly VPN with all the perks, highly ranked by PCMag and TechRadar. Use our link to save 86%, plus get 3 months free.
Detecting whether content is written by an AI programmatically is tough. AI-generated text often mimics human writing styles. It uses varied sentence structures, vocabulary, and context just like humans. AI can also learn from massive amounts of data, making its outputs increasingly sophisticated and harder to distinguish from human writing. AI-generated content can be tailored to specific styles or topics, adding to the challenge. Without clear markers or distinctive features, it’s difficult to create an algorithm that consistently and accurately identifies AI-generated text.
I doubt you noticed, but I wrote the previous paragraph with ChatGPT. That’s why outlawing AI at an organization level is as ridiculous as my boss at a major news org, long ago in 2001, outlawing the internet for reporters.
That said, writers have to be prepared to get dinged. Any writing that follows rote patterns and uses common expressions can potentially be AI-written. Like plagiarism checkers, AI detectors are blunt instruments that can pick up endless false positives simply because there is no new phrasing under the sun.
What should you do to avoid AI-written content? Well, you can basically tell your writers not to use it. This will result in writing that takes much longer and still includes good ol’ human errors.
But ask yourself why you’re worried about AI-generated content? Google doesn’t really care unless you’re really dumping absolute garbage on your site, replete with hallucinations and lies. Humans can rarely tell the difference, and if you’re using AI for entire articles you’re doing it wrong anyway. AI writing, as we’ve noted before, is best used as incidental filler. Need a paragraph on the history of the U.S. Congress? AI. Need a few grafs on the best coffeemakers? AI. Need an outline for a post on how to wash your dog during tick season? AI.
But if you keep a human in the loop you ensure that these paragraphs read well, flow fluidly into each other, and aren’t junk. If you fear AI the way we once feared the Internet, however, you’re for a bad time.
Leave a Reply