OpenAI Announces AI Detection Tools for Audiovisual Content

While the advancement of AI has opened up many possibilities in the field of technology, it is not exempt from being misused. Since the inception of AI, There have been concerns that the technology will contribute to a rise in misinformation online.

OpenAI has come forward with a new tool to combat the exploitation of artificial intelligence. The company has added a new tool to determine whether a picture was generated using its DALL-E AI image generator. It has also updated its watermarking techniques to be tamper-resistant in order to clearly identify and tag AI-generated audio content.

OpenAI has also officially joined the C2PA committee, which is responsible for establishing an open standard for labeling any content created using AI.

OpenAI outlined the new difficulties that have arisen with the introduction of AI-generated content in a blog post. “We believe it will be increasingly important for society as a whole to embrace new technology and standards that help people understand the tools used to create the content they find online,” the company stated, referring to the rise in generated audiovisual content.

Their image detection classifier predicts the likelihood that an image was produced by DALL-E. According to OpenAI, the classifier remains effective even if the image is cropped, compressed, or has its saturation altered. Although the tool has a 98 percent accuracy rate in identifying whether an image was created using DALL-E 3, it performs less well in identifying content that was created using other AI models, only identifying 5 to 10 percent of images created using other image generators, such as Midjourney.

There is still work to be done on both the audio watermarking signal and the image classifier. According to OpenAI, user feedback is necessary to evaluate the efficacy of its system. Using OpenAI’s research access platform, researchers and nonprofit journalism groups can examine the image detection classifier.