There were a few posts in the large CDG thread about it (which may be hard to find now) but the consensus seemed to be 223 fed ok running the bolt slow or fast because the 23 degree shoulder angle was more forgiving, but 223AI only fed ok if you ran the bolt fast, and if you ran it slow the 40 degree neck/shoulder junction had a tendency to get stuck on the edge of the chamber and stop the bolt. (This was the issue I had with the 223AI conversion on my AIAX, and probably why Enrique recommended I chamber it in 223 and not 223AI, lol)
I tried spinning one of my flat breech 223AI barrels from my Archimedes on my CDG just to see how things worked, it did not want to feed at all when running the bolt slow and only chambered about 40% when running the bolt fast... but that's with a non-conical breech. ARC says the conical breech is a must for 223 diameter cases, so my test was destined for failure.
IMO the 6 o'clock lug and lack of a full feed ramp in the CDG really isn't ideal for the small diameter 223 cases feeding from center feed AICS magazines as the small diameter 223 rounds ride low in the 6 o'clock lug relief and get presented to the chamber at a non-optimal angle, but since I don't have a 223 or 223AI barrel with a CDG conical breech take my little experiment with the flat breech 223AI barrel with a bit of skepticism.
Both my 223AIs are built on Archimedes actions specifically because I know they work perfect with 223 diameter cases and 223 AICS mags; the Archimedes has a full feed ramp for center feed AICS magazines and 223 & 223AI feeds great running the bolt slow or fast.