Wondering if there’s an easier way to detect ChatGPT-generated content? A watermark may open up that possibility.
What do you need to know?
A cryptographic watermark that could make it relatively easy to catch ChatGPT-generated content is moving its way in. ChatGPT recently introduced a way to create content instantly but is following up with a watermarking feature that’s sparking a note of fear among some. This is how ChatGPT works and how you may be able to slip past the watermarking feature.
ChatGPT is an innovative tool that SEOs and online publishers have a love/hate relationship with. Frenemies, if you will.
Some people love it because they’re discovering brand new ideas, formulating content briefs, and creating articles that are difficult to compose. Publishers aren’t feeling as optimistic about AI content taking over the internet and sparking the supposed “robots take over humans” idea that’s gaining traction more than ever. Watermarking is giving skeptics hope, however.
What is a cryptogenic watermark?
A watermark is a logo or text that is semi-transparent and is made to mark an image with its print, letting others know who the original author of the art is. You’ve probably seen plenty stamped on images, but video watermarking is more common now, as well. Watermarking text in ChatGPT infuses cryptography by embedding a selection of letters, words, and punctuation in the form of a unique code. Content created by AI is, let’s be honest, easy to spot. Human-made content follows a pattern more statistical in nature and is also easy to recognize. Sometimes, though, the lines are blurred because humans are smart and sophisticated.
The key to making AI content watermarking virtually undetectable is ensuring the distribution of words still has a unique and random appearance like regular AI-generated text.
You can’t currently use the ChatGPT watermarking feature, but it’s in the planning stage of development. It’s predicted to arrive soon- possibly in the final stage of ChatGPT feature rollouts.
