> Not all human subjects research requires informed consent — or even consent.
There have been multiple examples of this going horribly, horribly wrong. (Naturally, the worst examples were government-funded and ran during the Cold War.)
As a society, we have since concluded that at bare minimum, people should know they are being experimented on--and even that isn't enough to stop things from going badly.
This is why the IRB exists in the first place. A major part of its purpose is to prevent this sort of thing from causing undue harm, i.e., by forcing people with limited incomes to seek legal counsel because they believe they're about to be sued into the ground. One of the rules generally agreed upon for this is that experiments with human test subjects must inform those humans up-front what they're getting into.
To say that the response to this "can lead to science being much harder" is an ethically wrong defense. We know it makes certain kinds of research harder; that's the point. There are certain kinds of research that directly harm their subjects, and we don't do that to people. More than that, people have a right to decide whether they want to be involved in a study, as they may personally feel endangered by it (i.e., someone who has a PTSD response to being sued may not want to deal with being fake-sued).
To say that calls to dissolve the IRB "put too much responsibility on the institution" is flat-out false. This IS the IRB's responsibility--they approve or reject studies like this specifically to avoid ethical problems like this one. To claim that this isn't the IRB's responsibility is like claiming that it's not the responsibility of the law to revoke a driver's license when someone has been driving drunk, or that it's not the responsibility of the Food and Drug Administration to reject approval for foods that contain dangerous contaminants.