Social media users have complained that the AI image generator produces images that depict historical figures, such as America's founding fathers, as people of color.
In a post on social network X, Google said its AI feature was “used all over the world ,” but had created “inappropriate” products and the company was “immediately improving these descriptions.”
For now, the search giant will pause Gemini's human image generation feature and re-release an "improved" version as soon as possible.
The AI image generator was launched by Google in February through its AI platform Gemini, formerly Bard, at a time when the company is trying to catch up with OpenAI - the startup that owns ChatGPT, which just launched a new generative model capable of generating videos from user text prompts called Sora last week.
Sora is OpenAI's new generative AI model, which works similarly to the Dall-E image generator from the same house. Users simply prompt for video content and Sora returns a high-quality video clip. It can also create videos from still images, stretch videos, or fill empty frames.
While welcomed by AI enthusiasts, the new technology has also raised serious concerns about misinformation as major global political elections approach. According to data from machine learning company Clarity, the number of AI-generated deepfakes has increased 900% year-over-year.
To date, tech giants like Meta and Google have AI tools for creating videos, along with products from startups like Stability AI. Amazon has also released Create with Alexa, a model that specializes in creating short-form animated content for children based on prompts.
(According to CNBC)
Source
Comment (0)