Nit: I would be astonished to hear that any recent clinical trial involving placebos had taken place without all participants knowing that they might be given one. FDA guidelines on informed consent[0]: "Procedures related solely to research (for example, protocol-driven versus individualized dosing, randomized assignment to treatment, blinding of subject and investigator, and receipt of placebo if the study is placebo-controlled) must be explained. ... The description should also provide relevant information about any control used in the study. For example, whether the control is a medically recognized standard of care or is a placebo (including an explanation of what a placebo is)." Searching the web for "placebo IRB" turns up more detailed institutional guidelines; UCI's is one of the most in-depth[1].

I believe this is relevant because the author's comparison of placebos and virtue labelling hinges on whether or not the practices are deceptive. But the norms around informed consent mean that using placebos in clinical trials does not, IMO, involve deception. I'm not sure if there's a comparable mitigation to make virtue labelling not deceptive (can you imagine handing someone a consent form saying "I will sometimes lie to you about whether or not I think you are a good person, for your own good"?)

[0] https://www.fda.gov/RegulatoryInformation/Guidances/ucm40497...

[1] https://www.research.uci.edu/compliance/human-research-prote...