I tried using the newly released Research Assistant and specifically prompted it to generate an image explanation. In the steps of its response, it has a clearly labeled section for "Generating image", but it fails to show the image in the response. Instead, it says it cannot generate images in the current environment. Expanding the section about "Generating image" also does not show the image.
Here is a link to the thread: https://kagi.com/assistant/cff056c4-b4d5-4517-8f2c-bd10e331ab47
I expect an image to appear inline in the response. If there are technical limitations to doing this with Assistant, I think they should be documented.