(Aug 29): Alphabet Inc’s Google said on Wednesday that it would resume allowing some consumers to use its artificial intelligence (AI) models to generate images of people, after a scandal earlier this year prompted the company to suspend the feature. 

 In February, Google came under fire for its image-generation tool, which depicted historically inaccurate scenes, including incorrect racial depictions, when prompted to create images of people. Alphabet chief executive officer Sundar Pichai called the responses “completely unacceptable”, and the Mountain View, California-based company stopped accepting prompts for people while it worked to address the concerns raised.

Six months later, Google is ramping up once more. The company said it will begin rolling out image generation of people to English-language users who pay for its Gemini Advanced product, which offers access to some of the company’s most cutting-edge AI models. Businesses will also be able to use the tool, which will be available in the coming days.

Dave Citron, a senior director for product management at Google, wrote in a blog post that the company had made “significant progress” in generating depictions of people through Imagen 3, an AI model that underpins some of its products.

“We have worked to make technical improvements to the product, as well as improved evaluation sets, red-teaming exercises and clear product principles,” Citron wrote. Red teaming simulates using a product to find its potential harms before it reaches the public. “Of course, not every image Gemini creates will be perfect, but we will continue to listen to feedback from early access Gemini Advanced users as we keep improving.”

Citron said users would not be able to create photorealistic imagery, identifiable individuals, portrayals of minors or “excessively gory, violent or sexual scenes”. Google will gradually bring the feature to more users, he added. 

Uploaded by Tham Yek Lee