OpenAI has for more than a year had a tool to watermark text generated by the company’s AI assistant ChatGPT, according to information provided to The Wall Street Journal via The Verge. (OpenAI itself has confirmed that it’s been working on such a tool in a blog post.)
According to Journal, the watermarking should not affect the quality of the content ChatGPT creates — but OpenAI has chosen not to start using the tool because internally there are divided opinions about it.
The generative AI (genAI) company reportedly commissioned a survey that showed four times as many people are disposed to using tools that can identify AI content as opposed to them. At the same time, another survey found that 30% of ChatGPT users would use the service less if it contained watermarked content.
