Object Detection requires you to tag boxes around objects so you would still need to tag them. The scenario you are describing refers more to Custom Image Classification. This is a scenario that we are evaluating, let me know if you have a specific scenario in mind.
We are looking into ways to accelerate the tagging process through various means, but nothing will be available in the short-term.
since we create object "names" in ai builder we can have them like "folder" where we can import all images and instead of drawing boxes around them, all images that will be imported in that "folder" will be tagged with this object "name", there for no need to draw any box around it as the box can be "auto drawn" by image size.
p.s. i think its pretty cool that you guys are working on AI builder and I'm curios to see where it will go.