OpenAI Develops Accurate ChatGPT Content Detector but Withholds Release Over Watermarking Concerns

Aug 5, 2024, 05:09 AM

OpenAI has developed a highly accurate tool to detect AI-generated content, including text produced by ChatGPT. Despite its effectiveness, the company has decided to withhold the tool due to concerns about tampering and potential stigmatization for non-English users. OpenAI's internal debates have also delayed the release, with some fearing that watermarking could lead to false accusations of AI use. A survey found that 69% of ChatGPT users believe cheating detection technology could result in false accusations, and nearly 30% indicated they would use ChatGPT less if such measures were implemented. OpenAI is taking a deliberate approach to releasing these tools, aiming to balance transparency and user trust while addressing concerns about misinformation and authenticity. The company has reportedly had this system ready for years, according to TechCrunch and IndiaTodayTech.