OpenAI pulls plug on AI detection tool over low accuracy rate

Getting your Trinity Audio player ready...

OpenAI has shut down its artificial intelligence (AI) tool for identifying AI-generated texts over its “low rate of accuracy.”

The generative AI developers announced via a July 20 update that the tool will no longer be available to the public as it pulled down links directing users to the classifier. The AI classifier was developed in early 2023 and found mass application in the educational sector as professors relied on the tool to spot texts written using AI.

Barely six months after its launch, OpenAI disclosed that the tool failed to distinguish human-written texts from AI-generated texts accurately.

“As of July 20, 2023, the AI classifier is no longer available due to its low rate of accuracy,” Open AI said.

During the launch of the classifier in January, OpenAI warned that “it is impossible to reliably detect all AI-written texts,” noting that its offering had several flaws. At that time, the classifier labeled human-generated text as AI-generated 9% of the time, but it appears the tool became increasingly error-prone in the months after its release.

OpenAI noted that the classifier’s accuracy improved considerably as the text length increased. However, the inherent flaws forced the firm to pull the pin on the service, confirming that it will be probing innovative ways to identify AI-generated texts.

“We are working to incorporate feedback and are currently researching more effective provenance techniques for text, and have made a commitment to develop and deploy mechanisms that enable users to understand if audio or visual content is AI-generated,” the announcement read.

OpenAI has come under fire for ChatGPT’s diminishing abilities since its launch. According to Stanford research, ChatGPT’s ability to spot prime numbers has tumbled from 97.6% to 2.4%.

The platform’s ability to create new lines of code and audit capability in Web3 has also taken a significant hit in recent months, according to joint research from Stanford and UC Berkeley. Researchers say OpenAI’s ChatGPT-4 saw a massive decline in its math-solving and visual reasoning abilities.

Stiff competition in the AI space

The successes of ChatGPT have brought in new players in the generative AI space, with Microsoft (NASDAQ: MSFT), Google (NASDAQ: GOOGL), and Meta (NASDAQ: META) jostling for a piece of the pie. With OpenAI facing a barrage of regulatory troubles and class-action lawsuits, other firms are keen on encroaching on its turf with their new releases.

However, OpenAI shows no signs of slowing down as it announced a planned rollout of its ChatGPT for Android users. OpenAI faces stiff competition from Meta’s commercial release of Llama 2, a free and open-source generative AI platform.

Watch: AI, Blockchain, and secret to winning in technology