OpenAI’s New Tool: Detecting AI-Generated Images for Safer Content

Business View
4 Min Read

OpenAI often creates new tools to enhance the experience of users as well as to improve Artificial Intelligence (AI) techniques. The AI research and deployment company started this firm with a mission to ensure that AI benefits all of humanity. It became an important part of our usage, at least the Traffic Statistics of OpenAI. Speaking about the latest development of OpenAI, the company has recently lifted a veil from their new tool specifically designed to locate images created or generated by its text-to-image model, DALL- E3, the Microsoft-backed startup. 

In a released statement made on Tuesday, the startup shared this news officially. The development of the new tool came in the middle of increasing worries about the influence of AI-generated content in the global elections of this year. It is worth noting that many fake accounts use AI-generated images or videos to spread false and fictional information that disturbs the milieu, especially in tense situations. To stop such things, OpenAI is launching a new tool. As we mentioned above the new tool will easily detect images created by the use of AI or its text-to-image generator. 

The company said that before officially launching this the developers tested the tool to check whether it was working properly or not and the results clearly show that the tool correctly detected images that were created with the help of DALL-3 about 98 per cent of the time during the internal testing. This tool can also manage common modifications such as compression, cropping and saturation changes with minimal impact.  The other plans of OpenAI also include tamper-resistant watermarking to mark digital content for example audio, images or photos with a signal that should be hard to remove. 

Besides, the ChatGPT creator also shared the news of their collaboration with C2PA which is a renowned recognised standard for certifying digital content origins. With the partnership of the Steering Committee, the firm is also looking forward to contributing to the development of this standard, improving the transparency in digital content creation and editing processes. The ChatGPT creator has started accumulating C2PA metadata into images created by DALL-E 3 and plans to extend to Sora, a video-generating model. 

Moreover, as part of the efforts, the ChatGPT creator has also become the partner of an industry group that includes Adobe, Google, and Microsoft. This group aims to provide a standard that would help trace the root of different media. Last month, at the time of the ongoing general election held in India, fake videos of two Bollywood actors went viral in which they can be viewed criticizing Prime Minister Narendra Modi. These videos had accumulated millions of views in a short time and created a tense situation. 

However, it is not the first time that deepfake videos or AI-generated content have created a stir on social media as the trend or spread of AI-generated content is being increasingly used in India and elections elsewhere in the world including in Indonesia, the United States and Pakistan. The company states it is joining Microsoft in introducing a $2 million “societal resilience” fund in order to support the initiative for AI education. 

Read More:- Flydubai’s Summer Getaway: 10 New Flight Routes to Explore

Share This Article