Noise pollution is a bad thing. It takes quite a loud noise for people to start dropping dead the instant they hear it, there's plenty of research that shows that it is, in fact, not just a nuisance but causes health effects to man and animal alike. There are two big categories of car noises. Some noise comes from the engine and equipment and some comes from the interaction of the car's frame with the air or the tires with the road. Both have been reduced over time with various advances like mufflers and improved roadway design.
Hybrids and electric cars give even more opportunities for noise reduction because electric motors are nearly silent. Actually, even some standard car designs can get awfully quiet. People are starting to say that maybe hybrids are too quiet. There's clear evidence that, for cars in electric mode, under 20mph, they are much quieter than standard cars, to the point where the distance you can hear them from is dramatically reduced.
Now, in an ideal world, people are watching out for other road residents and noise need not be compulsory. This is why people are allowed to drive and bike with impaired hearing and blind people aren't barred from the sidewalks. But the reality is that if you've got hearing, it's a valuable source of added information because there are plenty of people not looking in the right direction on the roads. There is a reason why a lot of cyclist and pedestrian deaths involve cars.
The problem is congested situations like parking lots and stoplights. Remember, above 20mph there's enough road noise that you can hear a car coming. But a electric vehicle at 10mph not paying attention in a parking lot is a hazard. Thus, the National Highway Traffic Safety Administration found that hybrids are, in fact, more likely to hit pedestrians at these low speeds.
The proposed solution is to add noise, preferably only when the car is going at a low speed. The trick is finding a noise that is car-like such that your brain pegs it as car noise without needing to be told "Hey, that's a car" but also continues to help the overall noise problem.
Things are never this simple. The problem is that cyclists and pedestrians are disadvantaged users of the road. If I get hit by a car while biking, I have no guarantees that the car driver, even if it was clearly their fault, will suffer any repercussions for their act. Most of the time, the driver doesn't even bother to stop. I should add that there are now companies who would like to have a congressional mandate to force carmakers to buy their patented noisemakers. I'm sure that bribe...err... campaign contributions and bribe...err... industry consultation sessions are already going on.
So, I'm of the distinct opinion that the people who are actually making the decision have my best interests as a cyclist at heart.
Consider a driver with a aging hybrid. His noisemaker box fails. He may not necessarily realize that it's gone, nor is he going to try and have his box fixed quickly. It's not like his headlights, and after all, he's a careful driver. But suddenly, he's a lot more likely to run somebody over.
How do you solve this problem? Check it with the yearly inspection? What if it's only putting out half the noise, or working half the time? Will everybody check? What about places that presently don't require inspections on a regular basis? What about people who intentionally disable the box?
Unless it's a mechanical noisemaker attached to the electric motors in such a way that the motors fail to work without it, I can guarantee that some percentage, potentially an obnoxiously high percentage, of cars will have inoperative noisemakers. And part of the problem with a mechanical noisemaker on the motors is that you are back to square one as far as noise reduction, seeing no noise-reduction benefit.
Risk compensation is very popular. Daylight running lights, antilock brakes, bike helmets. It's fairly rational to assume that at least some drivers will assume that they are more audible to pedestrians than before, which is bad in the normal case and dangerous when the box has failed or the pedestrian is deaf.
I can see all sorts of ways that, by getting the details wrong (remember, Toyota just had to do a huge recall because of the driver-side carpet) that this can look like it's improving safety, benefit for a year or two, and then turn out making pedestrian and cyclist safety much worse. There may be a simple answer, perhaps not even by addressing the noise problem but addressing some aspect of the incompatibilities that are causing this type of accident, but you can bet that there's already enough money in the marketplace that it's a battle of lobbying, not hurting fewer people.