Microsoft is sharing access to the latest OpenAI text-to-image generator and claims to be taking steps to address safety concerns.
Microsoft is giving Bing Chat a major upgrade by making the text-to-image generator Dall-E 3 available to all users.
The tech giant said this AI model will improve the overall quality and detail of images, while giving improvements for human hands, faces and text in images.
Dall-E 3 is the latest text-to-image generator by ChatGPT-creator OpenAI, which was unveiled last month. These systems are able to create images based on user text descriptions. OpenAI claims this upgraded model is able to understand “significantly more nuance and detail” than its previous systems.
Microsoft’s announcement is significant as it means Bing Chat and Bing Image Creator users are getting access to this model before ChatGPT users.
“Since launching Bing Image Creator, more than 1bn images have been generated, helping inspire people’s creativity,” Microsoft said in a blogpost. “We’re excited for you to take your creativity even further.
Microsoft said Dall-E 3 can follow user prompts with more precision and reliability, to generate images that are photorealistic, creative and “artistic”.
“The images are not only visually appealing, but also logically consistent with the prompt,” Microsoft said. “The images can be uniquely styled with flair that meets your creativity.”
Addressing safety concerns
Concerns have been raised in the past around the use of text-to-image generators, due to their ability to create realistic images that could be used inappropriately or for misinformation.
Microsoft claims the company and OpenAI are addressing these concerns through a “content-moderation system”, which removes any images that contain nudity, violence, hate speech or illegal activities.
The tech giant also said all AI-generated images made with Bing Image Creator have an “invisible digital watermark”, which includes the time and date it was originally created and confirms that the image is AI-generated. Microsoft said this step adheres to specifications set by the Coalition for Content Provenance and Authenticity.
Earlier this year, the US government secured “voluntary commitments” from leading tech companies including Google, Meta and OpenAI to ensure the safe and transparent development of AI technologies. One of these commitments was to add watermarks to AI-generated content.
Last month, OpenAI said the images a user generates with Dall-E 3 are theirs and that they “don’t need our permission to reprint, sell or merchandise them”. But the legal realm of AI-generated images and copyright remains murky at best.
In August, a US court ruled that artwork generated by AI cannot be copyrighted. US district court judge Beryl Howell added that copyright has never been granted to work that was “absent any guiding human hand” and that human beings are an “essential part of a valid copyright claim”.
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.