OpenAI LLC at this time launched a cloud-based tool that may assist customers decide whether or not a bit of text was written by synthetic intelligence.
The tool, which is out there at no cost, is itself primarily based on an AI mannequin. It may possibly detect text generated by not solely OpenAI’s neural networks but in addition competing software program from different firms. Nevertheless, the startup cautioned that the tool presently has a variety of important accuracy limitations.
“In our evaluations on a ‘challenge set’ of English texts, our classifier correctly identifies 26% of AI-written text (true positives) as ‘likely AI-written,’ while incorrectly labeling human-written text as AI-written 9% of the time (false positives),” OpenAI researchers detailed in a weblog submit at this time.
OpenAI says that the tool is most adept at analyzing English-language texts comprising greater than 1,000 characters. It may possibly’t reliably consider text snippets which might be shorter or written in different languages. Furthermore, the tool isn’t but able to precisely detecting AI-generated software program code.
There are additionally different limitations to the tool’s accuracy. In accordance with OpenAI, AI-generated text might be modified to make detection tougher. The startup’s researchers added that “for inputs that are very different from text in our training set, the classifier is sometimes extremely confident in a wrong prediction.”
The coaching dataset that OpenAI used to develop the tool included a set of text prompts submitted by customers to its cloud-based AI fashions. In accordance with the corporate, its researchers entered every immediate into a number of neural networks to provide a sequence of machine-generated responses. The corporate additionally used text from different sources throughout the growth course of.
OpenAI’s efforts to develop software program for detecting AI-generated text date again a number of years. In 2020, the startup launched GPT-2 Output Detector, a tool for recognizing text generated by its GPT-2 neural community. The neural community is a predecessor of ChatGPT that was launched in 2019 and affords a extra restricted function set.
In accordance with OpenAI, the brand new tool it launched at this time is considerably more proficient at detecting text generated by lately developed AI fashions.
“While it is impossible to reliably detect all AI-written text, we believe good classifiers can inform mitigations for false claims that AI-generated text was written by a human: for example, running automated misinformation campaigns, using AI tools for academic dishonesty, and positioning an AI chatbot as a human,” OpenAI’s researchers defined.
The discharge of the tool comes every week after the startup raised a multibillion-dollar funding from Microsoft Corp. to advance its AI analysis. The funding consists of not solely capital but in addition cloud computing infrastructure. Beforehand, OpenAI raised $1 billion from Microsoft in 2019 and partnered with the corporate to construct a cloud-based supercomputer for its researchers.