The drones that are being developed are very stealthy, but it is important to understand that the purpose of stealth is not necessarily to make an aircraft completely invisible to radar. If that can be achieved then of course it’s great, but what really matters is preventing anti-air systems from being able to target the aircraft. Even if the enemy is able to see the aircraft, that’s of little value if their fire control radars can’t lock onto it.
It’s far easier to detect something than to target it. Second, as far as automation and someone’s contention that we don’t need the B-21 because unmanned drones can do its job, the B-21 reportedly is capable of operating in either manned or unmanned modes, depending upon mission requirements. The B-21 is far larger than any drone, so can carry a much larger payload, and where a single aircraft, or a few aircraft meet the mission needs, it would be foolish to send dozens of drones instead. For one thing, the B-21, like the B-2, will have global range, while much smaller drones would likely need to be refueled more often. Imagine one B-21 being capable of flying all the way to a target and only needing to refuel once on the way back, versus let’s say a half dozen drones (just to make up a number - I suspect it could be higher) that would all have to refuel on the way there and on the way back, possibly multiple times. That’s a huge difference in the logistics required, and the number of tankers loitering out there, loitering out there possibly being detected or becoming targets themselves.
We need both the B-21 and combat drones, but they fill very different roles. And until automation (AI) reaches a point where it can be trusted to react correctly to dynamically changing combat conditions, and to make the decision to release weapons, manned aircraft will still be at the center of any attack, with drones possibly serving as force multipliers under the control of the manned aircraft’s crew.