1. I very much like a suggestion of David Keith that we make a distinction between “processes” research and “sub-scale deployment” research.
2. Of course, this is a murky distinction – much process research is very sub-scale deployment.
3. I would argue that it would be a real contribution to characterize this distinction in clear terms with a goal of carving out the process work so it is either unobjectionable or already regulated. If these do not suffice, drawing up regulative guidelines would be very worthwhile. I taking the guiding idea to be this: process research can be reliably characterized ex ante as posing no risk.
4. When it comes to sub-scale deployment, if you think such deployment is plausible given the timeline David posed, I agree with a view espoused by David Morrow that consent is key. But I don’t think it is useful to use(as he suggests) a biomedical model for the following reasons:
i. The population of the world is not one person-like entity. So there is no one thing in which benefits and burdens can be balanced.
ii. A maximin approach (only pursue a course of action if you know ex ante that it will be at worst the best off all possible worst off outcomes) assumes we have knowledge of risks and we don’t.
iii. In hospital committees (and I have served on them) there are occasional such cases but they only proceed with terminal patients for which no existing treatment is available. Not only is the world not one patient it is not terminal!
5. You may say, “look, how can we make things worse than we are already with climate change”? The burden is on you then to show that you can’t. Not that you can’t on average, but that you can’t for at least some non-trivial parts of the world in some non-trivial ways. I don’t think you will be able to do that until and unless climate models improve to give confidence in predictions at a local level.
6. Without such confidence, I would urge a ban on sub-scale deployment for now.
7. This is no great sacrifice for advocates of such deployment, for all those who favor research have their hand full with the “process” agenda.