San Francisco, July 26: OpenAI, owned by Microsoft, has decided to discontinue its AI text-detection tool due to its “low rate of accuracy” in distinguishing human-written material from content generated by their AI chatbot, ChatGPT.
In a recent blog post, OpenAI acknowledged the limitations of the text-detection tool and expressed their commitment to develop more effective provenance techniques. The company is currently researching and working on an enhanced text version and has also pledged to do the same for audio and visual content produced by its Dall-E image generator.
The text-detection tool was initially introduced in January 2023 as part of OpenAI’s effort to address the issue of false claims in AI-generated text. However, it faced criticism for its limited and inaccurate performance. Users had to manually input at least 1,000 characters of text for the tool to classify it as AI or human-written content. The classifier’s accuracy was found to be just 26% in identifying AI-written text correctly, while it mislabeled human-written text as AI about 9% of the time.
In the midst of these challenges, OpenAI’s CEO, Sam Altman, has launched a new venture called ‘Worldcoin.’ This audacious eyeball-scanning cryptocurrency startup aims to distinguish humans from AI online, enable global democratic processes, and create greater economic opportunities. Worldcoin comprises a privacy-preserving digital identity (World ID) and, where regulations permit, a digital currency (WLD) that individuals receive simply for being human.
The World App, the first protocol-compatible wallet, is now available for users to download and reserve their share of the digital currency.
