OpenAI Will Launch AI Image Detection Tool Made DALL-E 3

Illustration of the image produced by DALL-E 3. (photo: x @openai)

JAKARFTA - OpenAI will launch a tool that can detect images created by their text-to-picture generator, DALL-E 3. This was announced by startup companies backed by Microsoft on Tuesday 7 May, amid growing concerns about the influence of content generated by artificial intelligence in this year's global elections.

The company says the tool correctly identifies images created by DALL-E 3 of about 98% of the time in internal testing and can handle general modifications such as compression, cutting, and saturation changes with minimal impact.

In addition, the creator of ChatGPT also plans to add water markers that are difficult to change to mark digital content such as photos or audio with signals that are difficult to remove.

As part of these efforts, OpenAI also joins an industry group that includes Google, Microsoft, and Adobe, as well as plans to provide standards that will help track the origins of various media.

In April, during India's ongoing general election, a fake video of two Bollywood actors seen criticizing Prime Minister Narendra Modi has gone viral online.

The spread of content generated by artificial intelligence and deepfakes is increasingly being used in India and in elections elsewhere in the world including in the United States, Pakistan, and Indonesia.

OpenAI said it would join Microsoft in launching a $2 million "community resilience fund" to support artificial intelligence education.