Probably. Perhaps it would be the best to handle such technology legally in the same way as possession of weapons is handled. Getting a license should be possible, but not trivial for every crazy person who wants one.
People can still get broken, but that’s not a very good arguments against use of such technology, since people are often pretty much broken by nature, or our artificial environment. If such tech gives them a chance to improve their situation, without causing too serious side-effects, why not?
What are “correctly educated agents”? I’d argue that a complete education should include the experimental use of such technologies on oneself!
While that may be true, this kind of reasoning could be applied to any kind of problem that we are faced with. Humans are too stupid. Let the AIs fix them. Well, we are faced with problems now, or very soon. There’s no time to wait for the saving grace of the AIs we are going to build.