OpenAI launched ChatGPT in 2022, and since then, many have been concerned about students using the AI to cheat on assignments. Even though AI detection tools exist, they aren’t the most reliable. However, OpenAI has now revealed that it’s developed a method to detect when someone uses ChatGPT with a 99.9 percent efficacy.
OpenAI’s AI detection will slightly alter each of the tokens and then leave a watermark, which its tool can spot. While this technology has reportedly been available for nearly a year, the company has had mixed reactions to it internally
The report says that some at the company think this tool will push away from ChatGPT’s user base, but using it would stand by its commitment to transparency. An OpenAI spokesperson also mentioned that this tool could “disproportionately affect groups such as non-native English speakers.” Despite this, most of the critical people at OpenAI support its AI detection tool’s propensity for good and shouldn’t be delayed.
While this is useful, there can be ways of going around this, such as using Google Translate to change the text to another language and then back to English. It’s unclear when the feature will officially be available, but educators will probably love this upcoming tool.
Header image credit: Shutterstock
MobileSyrup may earn a commission from purchases made via our links, which helps fund the journalism we provide free on our website. These links do not influence our editorial content. Support us here.
