Without any systematic data or evidence of a problem, or even a thoughtful analysis of costs and benefits, the application of our human subject review system within universities is overreaching at the same time that some risky experimentation on humans outside of universities is unregulated. This paper questions the purpose, feasibility and effectiveness of current IRB approaches to most "two people talking" situations and proposes scaling back our regulatory system to increase respect accorded it by researchers and its ability to protect human subjects of research from real, versus imagined harms. In too many cases, we are focusing upon form over ethical substance: counting what can be counted, rather than focusing instead on what counts. Some disciplines-oral history and journalism, for example-simply do not belong within the scope of Institutional Review Board jurisdiction. Others, such as survey research, informational interviews and informal interactions, call for a shift from centralized review to more departmentally-based (i.e., rooted in disciplinary ethics) oversight, and clearer guidelines on what requires advance review as opposed to provision of post-hoc complaint systems.