ScienceInsider has an excellent roundup of informed opinions here. I share the concerns of Harvard’s Marc Lipsitch:
“The fact that the global population is being put at risk by such experiments, to an appreciable but unknown degree, without being informed, much less consenting, is an ethical problem that has not been faced squarely.”
The real issue here is of tail risk. Accidents do happen in labs. In this case, the potential consequences of such an accident would be extreme. In a sense, this reminds me of the first test of the atomic bomb in New Mexico–the scientists involved weren’t completely sure that the thing wouldn’t ignite the entire world’s atmosphere. I suppose they went ahead because the value proposition in terms of ending the War was very attractive to the ultimate decision makers.