Google is updating its picture-creation technology to continue beating rivals.

Google today, at its I/O developer conference in Mountain View on Tuesday, unveiled Imagen 3, the third instalment in the company's generative AI model line called Imagen.

According to the head of DeepMind, an AI research division of Google, Imagen 3 understands the text prompts into images it translates better than its predecessor, Imagen 2, and has "more creative and detailed" generations, he said, adding that it produces "fewer distracting artifacts and errors."
"This is also our best model yet for rendering text, which has been a challenge for image generation models," Hassabis said.

To address concerns over creating deepfakes, Google says Imagen 3 will use SynthID-an approach developed by DeepMind to apply invisible, cryptographic watermarks to media.
Sign-ups for Imagen 3 in private preview are available in Google's ImageFX tool, and Google says the model will "come soon" to devs and corporate customers using Vertex AI, Google's enterprise generative AI development platform.

Google usually doesn't talk much about the source of data used to train its AI models — and this time was no different. There is a reason for that. Much of the training data comes from public sites, repositories, and datasets all around the web. And some of that training data, specifically the copyrighted data scraped without permission from content creators, is a source of IP-related lawsuits.

The control that Google offers its web publisher controls enables webmasters to prevent the company from scraping data, including photos and videos, from their websites. But Google does not offer an "opt out" tool, and, unlike some of its competitors, the company hasn't vowed to compensate rights holders for their (in some cases unknowing) contributions to the training datasets.

Surprisingly, there's also an element of secrecy at Google. It does lack the transparency. Such disappoints, especially a firm which has the type of wherewithal such as Google.