That was the next thing I thought of!
But even more of a reason for law enforcement to oppose it....revenue goes away. No more speeding, running red lights, parking tickets means a massive reduction in revenue. It won't be the police specifically....but which ever government branch gets this money will not like that very much.
It'll be difficult to build much momentum around that though, because in my experience for every traffic cop out there that loves to give out speeding tickets, there's a couple that don't and at least one that has nightmares about the last time a driver challenged the laws of kinetic energy transfer and lost big time.
Conversely, if municipalities in the long run have to spend less money on managing traffic, traffic accident disruptions, etc, it could end up at least washing out the overall expense. And automated driving significantly increases the carrying capacity of existing roads, reducing the need to add traffic capacity to highways and city streets.
Its really difficult to predict how it will all play out. For example, when they began replacing incandescent light bulbs with LED lights in traffic lights, the logic was that because they used less electricity and lasted longer they would end up costing less to use overall. And that was true - unless you did this in an area where it snows a lot. The old incandescents were so hot they would melt any snow that landed in the lightbulb shield. The LEDs were cold, allowed snow to build up in the traffic light assembly, and could eventually allow snow to block the bulb entirely. So it became necessary to send road crews to clear the lights in some cases, which increased costs until they retrofitted lights in those areas with snow deflectors. But even with the deflectors, the LED lights still end up costing a lot less to operate overall. To really get the net impact of automated driving, you'd have to sum up a very large number of collateral effects.