Drawbacks aren't "drawbacks". You have to look at things synthetically. There are always drawbacks to any course of action, but that might not mean it's the correct one.
Re: Yankee: I think that the command and control that Trident suggested is wrong, simply because the Su-27PU was built for the Soviet doctrine, which emphasized ground-control interception and would require a command Su-27 to help transition the Soviet Air Force's doctrine.
===
Before I comment further, let me put out some basic facts.
-A J-20 is never going to be as capable as a KJ-200 if we ignore its speed and stealth advantages. A KJ-200, not to mention a KJ-3000, is carrying a larger AESA system that is at least L-band. The J-20 can also go L-band, but it will be limited by its aperture size. On a KJ-3000, you can have a VHF or UHF-band AESA with counterstealth capability that can detect a stealth aircraft at relatively long ranges.
-A J-20 dedicated to command roles is going to be compromised compared to a normal J-20 due to its twin-seat nature. It will have a longer effective range for enemy missiles due to its inferior kinematics, and will likely sacrifice some of its stealth. Moreover, since it is a force multiplier, it is going to be a priority target, and since its survivability has been reduced, it makes it easier for its opponents to neutralize it quickly, making it as though it was never here at all.
====
In the PLAAF's case, a command J-20 leading J-20s would be a step backwards. The only realistic implementation of a command J-20 would be a command J-20 supporting J-11s, J-10s, and JH-7s. The command J-20 would have tremendously increased survivability compared to the aircraft escorting it, making it nonsensical to shoot down the command J-20 before the escort fighters are knocked out. On the other hand, the J-20 would complement the 4th generation fighters it escorts by having greatly superior sensors as well as obfuscating the exact strength of the fighter group. An attacking force could expect a group of 4th generation aircraft to be easy prey, destroyable by its own 4th generation fighters, only to be surprised by a hidden 5th generation supporting element with situational awareness superiority.
For twin-seater J-20s in J-20 formations, on the other hand, the best approach would instead be an EW J-20 as opposed to a "command" J-20. An EW J-20 can compromise its offensive role and go further into its supporting role by emphasizing its ability to jam enemy sensors and missiles. Likewise, since the EW J-20 is not a command J-20, with the flight / squadron leader flying a conventional J-20, the loss of the EW J-20 degrades the ability of the J-20 flight / squadron, but not to the same degree as if the command J-20 were knocked out instead. The EW J-20 could even serve as "vice flight/squadron leader" and lead as a back-up should the J-20 flight leader be knocked out, increasing the robustness and resilience of the flight. And since an EW plane is always an attractive target, it protects the command-and-control system of the J-20 flight by taking hits before the actual flight / squadron leader gets knocked out.
====
For the twin-seater approach, the best model for comparison would be the Su-27. The Su-27 is not an aircraft, but a family of aircraft whose capabilities diverged and evolved radically. The Su-27 is a basic air superiority fighter. The Su-30 is a variant with enhanced strike capability as well as twin seats. The Su-33 is an adaptation of the Su-27 for carriers. The Su-34 is an adaptation of the Su-34 to light bombing roles. The Su-35 is the final evolution of the Su-27 line, taking up capabilities of the Su-27, Su-30, but not the Su-33 and Su-34 in a highly modernized and maneuverable package.
By going twin-seater, the point isn't that the J-20 is going to be doing ALL of these roles. That's essentially the mistake the Americans made with the F-35, that the variants are for different branches, but are intended as all-in-ones. By going twin-seater, the J-20 is extending its development path. Off a twin-seater J-20, you now have an easier time developing a strike variant of the J-20 (main problem is the bay length, not the depth, as I've mentioned. Changing the weapons bay configurations to something more similar to the F-35, with longer side bays, repositioned rear wheels, and smaller main bay might suffice for longer strike missiles) as well as a light bomber version. More importantly, it also enables the development of dedicated EW variants.
====
To summarize:
-If a J-20 twin seater ends up being a command aircraft, its best place is with 4th gens, not 5th gens, because it'll be compromised and a target with 5th gens.
-The J-20 twin seater is more interesting as a transitional aircraft to dedicated variants a la the Su-27 that perform a specific role.