Who exactly needs to get approval from an institutional review board (IRB)?

Scope of IRB Requirements (Law vs. Practice)

• Several commenters note that US federal statute (42 USC 289) ties IRB requirements to research funded under specific federal chapters; in principle, purely private research is outside that scope.
• Others emphasize that universities, hospitals, and pharma routinely extend IRB requirements to all human-subjects work, regardless of funding source, mainly for risk management.
• A New York state definition of “human research” is discussed; there’s disagreement whether something as informal as friends experimenting with diets clearly fits that definition.

What Actually Needs IRB Review / “Exempt” Work

• Many point out broad exemption categories (e.g., “benign behavioral interventions,” some retrospective chart reviews, secondary use of de‑identified data).
• However, multiple researchers (radiology, EEG, etc.) say that even “IRB‑exempt” studies usually require a submission and formal exemption decision; the process can still take months.
• There’s sharp disagreement over claims that deep-learning on clinical MRIs “doesn’t require an IRB”: others insist it does, with consent/notification commonly waived but still formally reviewed, especially given re‑identification risk.
• The notion of “theoretical risk” is criticized as so broad that nearly all human research qualifies.

Risk, Liability, and Institutional Motives

• IRBs are framed by many as primarily tools for institutional liability and reputation management (similar to boards or safety committees), not pure ethics.
• Several comments stress that even if government enforcement is unlikely, civil suits, funding loss, and institutional sanctions are very real. Waivers may not protect individuals if something goes badly wrong.

Corporate A/B Testing vs Academic Research

• Commenters contrast strict academic IRB processes with largely unregulated corporate A/B testing and engagement-optimization experiments on users.
• Explanation offered: institutions receiving federal research funds must comply with the “Common Rule,” while most corporate experiments do not fall under those rules unless involving regulated medical products or federal contracts.

Burden, Misalignment, and Weaponization Concerns

• Many researchers describe IRBs as slow, opaque, and often misaligned with actual risk (e.g., weeks of scrutiny for low‑risk surveys vs. seemingly limited safeguards in physically risky studies).
• Some argue IRBs can be “weaponized” to block or police disfavored research; others respond that the core problem in such cases is bad actors, not the existence of oversight.
• There is debate over whether IRBs meaningfully prevent harm or mostly smother benign research, with references to experiences in the developing world.

Rule of Law, Vagueness, and Enforcement Reality

• Several comments generalize from IRBs to broader issues: vague laws, expansive regulations, and the gap between “what’s written” and what is ever enforced.
• Engineers especially are seen as prone to over‑focusing on text rather than prosecutorial discretion and political incentives.
• Some argue that because enforcement is selective and precedent sparse, the real constraint is fear and institutional caution more than clear, democratically legible rules.

Reactions to the Article Itself

• Some praise the author for deeply engaging with regulatory details.
• Others think the piece overthinks edge cases, omits clear statutory limits, or uses a “cute” rhetorical style that feels biased against IRBs.
• There’s modest agreement that IRB rules are confusing, but disagreement over how much that actually matters in day‑to‑day research.