OpenAI has had a system for watermarking ChatGPT-created text and a tool to detect the watermark ready for about a year, reports The Wall Street Journal. But the company is divided internally over whether to release it. On one hand, it seems like the responsible thing to do; on the other, it could hurt its bottom line.
OpenAI’s watermarking is described as adjusting how the model predicts the most likely words and phrases that will follow previous ones, creating a detectable pattern. (That’s a simplification, but you can check out Google’s more in-depth explanation for Gemini’s text watermarking for more).
The company apparently found this to be “99.9% effective” for making AI text detectable when there’s enough of it — a potential boon for teachers trying to deter students from turning over writing assignments to AI — while not affecting the quality of its chatbot’s text output. In a survey the company commissioned, “people worldwide supported the idea of an AI detection tool by a margin of four to one,” the Journal writes.
But it seems OpenAI is worried that using watermarking could turn off surveyed ChatGPT users, almost 30 percent of whom evidently told the company that they’d use the software less if watermarking was implemented.
Some staffers had other concerns, such as that watermarking could be easily thwarted using tricks like bouncing the text back and forth between languages with Google translate or making ChatGPT add emoji and then deleting them afterward, according to the Journal.
Despite that, employees still reportedly feel that the approach is effective. In light of nagging user sentiments, though, the article says some suggested trying methods that are “potentially less controversial among users but unproven.” Something is better than nothing, I suppose.