A while ago Ozy wrote a post on the curb cut effect. They claim that when people without disabilities use things intended to benefit people with disabilities it's generally positive. For example, curb cuts and closed captioning were initially created for people in wheelchairs and people with hearing disabilities, but there's no harm in other people using them and having more people that benefit from them makes them politically sturdier. For example, they argue that:

Nondisabled people using wheelchairs does nothing but create a larger pro-wheelchair demographic, which benefits disabled wheelchair users.

But consider how buses handle wheelchairs. The driver positions the bus next to the curb and folds down a ramp. The passenger rides up the ramp, either they or the driver fold up a set of seats, the passenger positions their chair where the seats were, and again either they or the driver set the straps to hold the chair still. There are two spots for wheelchairs on each bus, and when a spot is in use three standard seats are unavailable. If lots of people who didn't need wheelchairs started using them, at first we would have a dramatically larger incidence of problems where buses already have their two spots occupied and don't have room for a third wheelchair-using passenger. Additionally, each bus could seat four fewer non-wheelchair-using passengers. Then maybe we redesign and replace the buses to allow for more wheelchairs, but this is expensive and capacity goes down even more which means we need more buses and more drivers.

Alernatively, consider someone who hates cilantro telling a restaurant that they're allergic to it. This does reduce their chances of getting food that tastes terrible, but it also requires the restaurant to put much more work into preparing it (pdf). The more people describe their preferences as allergies, the more time the restaurant will need to take with each order and the more expensive it will be for everyone.

Or imagine you're learning sign language, and you're going to a performance at a venue that provides intepreters when one is requested. If you ask for one so you can practice your ASL, you're making this service more expensive when considered on a per-person-who-needs-it basis.

This isn't to say that people who need wheelchairs shouldn't take buses, people with food allergies shouldn't eat out, or Deaf people shouldn't go to shows! It's really great that we now have low-floor buses and restaurants that handle allergies, because food and transportaion are things everyone should have, and it would be great if more venues offered ASL interpretation on request. There is a shared cost of using these accommodations and it's generally well worth paying in the case of people who need it.

Ozy writes:

If an accommodation helps you and you can get it without proving you're disabled (i.e. as you must to get a service dog), you should use it.

You don't have to prove you're disabled to bring a wheelchair on the bus or tell a restaurant that you're allergic to something, but if lots of people followed Ozy's recommendation I think that's the most likely outcome. Instead of the current low-friction honor system way of handling eligibility we would need a system for proving to the government that you did actually require the accommodation. That system would have some error rate, where not everyone who actually needs the accommodation would end up with a permit, and it would be a big hassle even for the people who successfully get their permits.

Some accommodations, like closed captioning or color communication badges (pdf), make things better the more people who use them regardless of disability status, while others, like wheelchair spots on buses, don't. If you're considering starting using an accommodation that is typically seen as being only for people who need it, try to figure out first whether it's something that imposes costs on others.