Wouldn’t work. The “rubbernecking” factor is overreated.
I think the OP is misinterpreting the cause of the jam.
Once the initial slowdown starts from either looking at the original accident, or just by having people go to their brakes in a small moment of panic, THE JAM EXISTS. The slowdown is not caused by successive rubberneckers after a point, but by the ORIGINAL rubberneckers.
Jams persist even after accidents have been moved. You’ve surely seen that. When traffic is moving along at 70 mph and there is a big slowdown, the JAM has begun and it’s going to survive.
Imagine if you were somehow to instantly remove the “attraction”. Or, put up your barriers. What happens? The OP supposes that the jam just frees itself all at once because there’s nothing to look at.
In reality, you have the first car accelerate; when the second car sees it, and gets his room, he accelerates; when the third car sees that, and gets space. . .well, you get the point. If you have 20 cars in the jam, the last 15 cars are still going 20 when the first 5 cars are accelerating away.
However, the whole time these cars are accelerating from 20 mph to 70 mph, additional cars travelling 70 are running into the back of the JAM which is still moving at 20 mph; if there is enough volume on the road, even with no attraction, the jam is not relieved, but can still propagate backwards.
Think about what you do when you see an accident. You’re sitting in a slow moving patch of traffic behind me. You’re not rubbernecking. I’m not rubbernecking; I’m just going the speed of the guy in front of me, and when he finally accelerates, I accelerate a second later. But, he’s not rubbernecking either, he’s going the speed of the guy in front of him.
After the initial rubberneckers, no one has to rubberneck at all for the jam to persist. The jam doesn’t clear until volume drops enough so that there is enough space between the last car doing 20 and the next car doing 70 so that by the time the guy doing 70 has reached the jam, the guy doing 20 has accelerated sufficiently.