Resistance Is...Reversible?
While people generally don't talk about bacteria losing antibiotic resistance, it does happen, and for a pretty obvious reason: the biochemical tools which confer resistance come at a metabolic/reliability/efficiency cost, or the bacteria would have had the resistance to begin with. In general, we can assume that bacteria are not adept at dealing with silver toxicity because silver does not occur often enough in their favorite environments. When you take the bacteria out of a silver-concentrated environment, there is no longer pressure to take extra measures to deal with it, and the resistance can be lost.
A Pound of Flesh
There are many ways to attack microbes, and even more ways for them to respond. The CDC provides a nice overview of some strategies.
Take fluoroquinolones, for instance. They kill bacteria by suppressing the enzymes which uncoil DNA. As you can imagine, this makes for a Very Bad DayTM for the bacteria. If you're a bacterium, and you end up with some ciproflaxin inside your membrane, the most obvious thing to do is get rid of it. Some bacteria, like P. aeruginosa, which are resistant to ciproflaxin do just that: they build pumps in their membranes to pump it out. But as any engineer will tell you, adding new pumps to a system costs both space and energy which could have been used for other things (like finding food, reproducing, etc.), so you wouldn't do this unless you absolutely had to.
How do we know that they actually lose resistance? Well, we've observed it happening! We literally know which mutations and gene transfers are necessary to confer various resistances for P. aeruginosa, because wild strains are both resistant and susceptible, depending on their exposure to antibiotics (or laboratory conditions).
Gray Goo
Be glad that biology tends to find the most efficient solutions. This energy minimization tendency might be characterized as a kind of laziness, but it's really an optimal strategy for allocating your energy where it gives you the most value. For a microbe, that would be finding food and reproducing. That means that once the lab-grown freaks escape into the wild, whatever super-powers they gained from our mad scientist experiments are likely to be lost over time, once we stop bathing them in mutagenic toxins.
Unlike nanotechnology-gone-wild, we actually don't have so much to fear from lab-induced antibiotic resistance. It is the pernicious and casual use of antibiotics in our daily lives that will kill us in the end. These flow down our drains and into our streams and rivers and evolve those critters night and day, virtually guaranteeing that we will encounter resistant types in the wild.