I'm currently working on my CITI ethical human research certification. Given its dense nature my mind started to wander and I had the realization that all the protocols and ramifications I've so far read have only to do with institutional funding and recognition. In theory what if there was an eccentric billionaire who was self-funded and published (or didn't care about becoming published), with a personal interest in human behavioral studies. Could they circumvent all the regulations and create an unethical study without punishment? I suppose research participants could sue, but if said hypothetical researcher didn't have to follow any IRB guidelines couldn't they add a clause to their consent waivers in which participants waive their right to financial recourse? This isn't a pressing question, just curious and likely missing the obvious. Thank you.
In order for sanctions to be applied for unethical and/or illegal research practice, it has to come to the attention of some authority who is willing and able to act. That might be one or more of the subjects of the study, if human, or an observer, or, well, anyone who knows of it.
But it is possible, in theory, for evil to occur outside the range of light. If no one knows of it, punishment isn't likely to occur.
But if the "research" is published, even by a private individual, then, certainly, people will know of it and if it also breaks civil or criminal law, action is possible assuming anyone reads it. Reputable publishers ask for evidence that the research was vetted by an independent ethics body, such as an IRB. Disreputable publishers might not.
Note that an IRB is a protection for the institution as well as for subjects. Likewise, publishers want to be protected against publishing harmful material.
Their have been cases, of course, in which certain oppressors have been killed by those they have harmed. The law may or may not punish the victim/punisher in such a case.
"In The United States" is not well defined. Do you mean US journals? US publishers? Or US-based authors? I think the answer to your question is that, from the point of submitting a manuscript for review to the point of publishing, there is barely any ethical oversight.
Recall the controversy over the article in which researchers rated the attractiveness of women and correlated their findings with prevalence of endometriosis. This should have been a fairly obvious violation of the Belmont Report. In addition to not getting paid, most reviewers don't actually understand their scope - this partly explains why the process is so wildly variable. Somehow this paper was reviewed and published. The ethical oversight of published research, then, seems to be entirely post hoc and really boils down to the community raising a unified voice. While ethics and populism rarely intersect, in this case I'm glad this dumb paper was stricken.
The major issuances on ethical research are the Declaration of Helsinki and Nuremberg Code which largely apply to interventional and prospective research, and concern the manner and nature of data that are acquired for research purposes. In addition, there is the Belmont Report, which addresses messaging, utility, and value of research. There are other regional laws that concern data privacy, data sharing, and medical conduct like Interantional Council on Harmonization, the code of federal regulations, and good X practice. These lead to the formation of independent review committees like data monitoring committees (DMC), internal review boards (IRB), and steering committees (SCs). Typically, IRBs review and approve protocols and research proposals and are housed within institutions that individually contribute data, DMCs monitor ongoing clinical research and are independent consultants hired by pharma companies or medical universities, and SCs will review manuscripts and are research staff within data coordinating centers.
I have personally voiced complaints against several SCs that believe open data don't require ethical oversight. I wasn't taken seriously, but I think it's an ongoing issue.
The most famous example of a "medical study" that we now believe would have never started is the Tuskegee Syphilis Study. But there are other significant studies that fell through the cracks. Examples include the Stanford Prison Experiment, the Milgram Experiment. These all underwent some internal approval by ethics boards (IIRC). Lastly, another very recent example concerns a study of regional alcohol use that singled out Barrow, Alaska and led to a massive decline in the "social credit" of this region https://psycnet.apa.org/record/1989-36833-001
You question whether self-funded and published research receives ethical oversight. The assumption here is that funded research is more critical received, and that unpublished research does not have penetrance in the community. Pause. Recall that during the era of COVID, "unpublished" or "preprint" papers were being cited prominently in media. Furthermore, if the authors declare conflict of interest (COI) on the paper, what additional effort if any do you believe the journal takes to evaluate the credibility of the authors? If authors lie about COI there are some watchdog groups such as Retraction Watch that will seek these issues. Researchers found guilty of misconduct can be blacklisted from participation in research and consulting.