One way to address this is to create different standards of evaluation and to adjust the legal liabilities accordingly. What I mean (very roughly) is something

One way to address this is to create different standards of evaluation and to adjust the legal liabilities accordingly. What I mean (very roughly) is something like: Untested => you are on your own, don't imagine you can sue anyone if something goes wrong, this is the default Conditional => some basic efficacy/safety protocols have been evaluated, some limited liability for the drug companies, Approved => full suite of evaluation done, liability still capped but at a much higher level then Conditional There would still be legal safeguards around fraudulent or negligent practices of course that wouldn't be constrained by the capped liability I suggested above.