That photo above? I didn't take it while writing this article in a coffee shop. I created it sitting at my home office desk, using Microsoft's new AI image generator, MAI-Image-1.
In AI vocabulary, the new tool is referred to as a "text-to-image model," meaning you write a prompt, and then the AI tool delivers an image. The prompt I used for this image was, "Create a picture of two people talking and having coffee at a coffee shop."
But I could easily have used the image below instead, since the tool spat out two for me to choose from.
Not real people, not real coffee. Microsoft
Microsoft's new AI image generator has not been officially released in the marketplace yet. You can access it by visiting LMArena -- a community platform where people test out different AI models. LMArena also has a ranking system: people compare two AI models side by side and vote for the one with the best AI output.
At the time this article was published, Microsoft's MAI-Image-1 was ranked in the top 10 of the LMArena rankings.
Don't miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.
In its blog post, Microsoft said the tool is designed to avoid "repetitive or generically stylized outputs." The AI team added that they are getting feedback from creative professionals.
"MAI-Image-1 excels at generating photorealistic imagery, like lighting (e.g., bounce light, reflections), landscapes, and much more," the post said. "Its combination of speed and quality means users can get their ideas on screen faster, iterate through them quickly, and then transfer their work to other tools to continue refining."
AI image generation is not new to Microsoft products. You can also create images with Microsoft Copilot or Bing Image Creator, but those are powered by AI models from OpenAI.
... continue reading