Collaboration does not necessarily mean direct control. Either way controlling a drone/CCA is not pilot directly flying them like an RC plane. It is more like RTS where you give it a task and watch it perform. But doing so still requires screen time so having someone other than the pilot handling it is optimal.
I want to address that bolded part again, because what you are talking about is entirely dependent on the task you are giving to the drone and the extent of automation that your collaborating manned aircraft has in context of other tasks.
If your drone is simply being told to fly in formation alongside a manned aircraft as a reference, or doing a preset patrol over a designated target area, etc, then that is something easier to automate than having a drone actively search among a variety of ground based targets to identify a specific vehicle type or building to strike.
The sophistication of the task in turn relates to what the onboard command capability your collaborating manned aircraft has -- which is dependent on crew size yes, but it also depends on the automation it has, the more of the aforementioned task it can do (whether it's simple or complex). And naturally if your manned aircraft itself has to focus on a complex tactical situation itself, then that will reduce the net capability to effectively command the drone, in relation to the sophistication, and automation.
A simple expression of it might be:
Drone command capability for a given task ∝ (Automation x Crew Size x Drone Task Simplicity)/(Manned aircraft task complexity)
In this case, given they're doing a simple formation flight (aka high Drone Task Simplicity), and considering the Automation capability of a J-20 likely is not inferior to a J-16D, and given the J-20 itself is just flying in a formation (aka relatively low Manned aircraft task complexity), then it should be very plausible for a single seat J-20 to command this drone for this specific task, even though it is "only" one single crew member versus two in the J-16D.
Putting it another way, I wouldn't be surprised if a single-seat vanilla J-20 can probably command many more drones for this specific task (flying in formation) than just a single GJ-11... therefore we should stop thinking about things like "drone command" as being something that is relegated to twin seat aircraft.
If one held all else equal (Automation, Drone Task Simplicity, Manned aircraft task complexity), then yes the Crew Size would positively correlate with "Drone command capability" and a twin seater aircraft variant has higher command capability than a single seat aircraft variant.
But for something as simple as flying in a formation in a non-complex tactical situation, that is a command capability which a modern single seat manned-aircraft should be able to do with ease. The benefit of a second crew member occurs if your Drone Task Simplicity is low (i.e.: you're asking your drones to do more sophisticated missions, or if you are managing a very large number of drones, or both), or if your manned aircraft is simultaneously engaged in high level/complex tasks (such as actively engaging in organic aerial warfare, EW, near the FEBA etc)