Published on November 8, 2023, 4:13 pm
Artificial intelligence (AI) has been a topic of concern lately, particularly when it comes to AI-generated images. People worry that these images could be used to spread misinformation and deceive the public. OpenAI, the creator of ChatGPT, is now developing a tool that claims to have 99% accuracy in detecting AI-generated images.
According to Bloomberg, this tool by OpenAI aims to identify user-made pictures produced by its Dall-E 3 image generator. At the Wall Street Journal’s Tech Live event, Mira Murati, the chief technology officer at OpenAI, stated that the tool is “99% reliable.” Although it is still being tested internally, there is no official release date announced yet.
If OpenAI’s claims hold true, this tool could provide reassurance to the public by distinguishing between genuine and AI-generated images. However, the company has not disclosed how exactly this tool will alert people to AI images — whether through watermarks, text warnings, or another method.
It’s important to note that this particular tool is designed specifically for Dall-E images and may not detect fakes generated by rival services like Midjourney, Stable Diffusion, and Adobe Firefly. Despite this limitation, any tool capable of highlighting fake images can have a positive impact.
OpenAI has previously released tools aimed at identifying content produced by its own chatbots and generators. However, one such tool meant to detect text created by ChatGPT was withdrawn due to its high level of inaccuracy.
In addition to discussing the image-detection tool, OpenAI also addressed efforts to reduce ChatGPT’s tendency to “hallucinate,” or generate nonsensical and fictional information. Murati mentioned significant progress with GPT-4 in resolving this issue but acknowledged that further work on GPT-5 is already underway.
In March 2023, tech leaders signed an open letter urging OpenAI to pause the development of any AI more powerful than GPT-4, highlighting potential risks to society and humanity. However, it seems that OpenAI has not heeded this request, as development continues at a rapid pace despite these concerns.
While there is uncertainty surrounding OpenAI’s new image-detection tool and its effectiveness compared to the previously canceled tool, what remains clear is that OpenAI is committed to pushing the boundaries of AI technology.
With ChatGPT’s ability to engage in natural language conversations and the introduction of GPT-4, which showcases even more advanced capabilities, OpenAI is making significant strides in the field of artificial intelligence. As users navigate this evolving landscape, understanding these powerful AI systems becomes crucial.