OpenAI will release a tool for identifying photographs made by DALL-E 3.

In response to mounting concerns about the impact of AI-generated content on this year’s global elections, Microsoft-backed OpenAI announced on Tuesday that it will release a tool that can recognize photographs created by its text-to-image generator, DALL-E 3, according to Reuters.

During internal testing, the tool discovered images taken by DALL-E 3 around 98% of the time. It can also tolerate common tweaks like compression, cropping, and saturation changes with little to no affect.

ChatGPT’s author also wants to integrate tamper-resistant watermarking, which will be used to designate digital items such as audio or photographs with a difficult-to-remove signal.

OpenAI is planning a standard that will aid in identifying the sources of diverse media, and as part of its efforts, it has joined an industry body that includes Google, Microsoft, and Adobe.

Fake films of two Bollywood actors denouncing Prime Minister Narendra Modi became viral in April, coinciding with India’s current general election.

Deepfake and AI-generated content are increasingly being used in elections around the world, notably in Indonesia, Pakistan, the United States, and India.

To promote AI education, OpenAI and Microsoft have announced the creation of a $2 million “societal resilience” fund.

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button