May 07, 2024
OpenAI, an American artificial intelligence organisation, is developing a tool to recognise photos made by its text-to-image generator DALL-E 3
The Microsoft-backed firm on Tuesday said that amid growing concerns about the impact of AI-generated material in this year's worldwide elections.
The company confirmed that the programme properly detected DALL-E 3 photos 98% of the time during internal testing and can handle standard alterations like compression, cropping, and saturation changes with minimum damage.
Moreover, the ChatGPT inventor also intends to include tamper-resistant watermarking to label digital assets, such as photographs or audio, with a signal that will be difficult to erase.
In addition to its efforts, OpenAI has joined an industry consortium that includes Google, Microsoft, and Adobe, and hopes to develop a standard to assist track the provenance of various media.
In April, during India's current general election, bogus films of two Bollywood stars criticising Prime Minister Narendra Modi became popular online.
AI-generated material and deepfakes are increasingly being employed in India and other countries, including the United States, Pakistan, and Indonesia, during elections.
OpenAI announced that it is partnering with Microsoft to develop a $2 million "societal resilience" fund to assist AI education.
