Over at LawfareBen Wittes shrugs off Jeff Stein’s important question,“What Will U.S. Forces Do With ISIS Prisoners?” Ben’s reaction, I believe, minimizes several important legal and policy issues that would arise if the United States military, or its partners, were to hold detainees.

Ben writes, “As long as U.S. force is largely a matter of air power and Iraqi and Syrian rebel forces are working the ground, Stein’s question will answer itself: U.S. forces won’t have ISIS prisoners, though the U.S. will have to look the other way as abuses of captives take place that our forces would not tolerate.”

His first point seems correct: airstrikes alone won’t lead directly to the holding U.S. detainees. His second point – that the U.S. “will have to look the other way” – may not be so simple, especially if Congress is authorizing funds that fall under the Leahy Law.

This law prohibits the provision of certain assistance and training to foreign security units where “credible information” suggests they have committed gross violations of human rights. As Steve Rickard noted in his testimony before Congress, the Leahy Law doesn’t require the U.S. to have pristine allies, but they do cover a short list of the most abhorrent human rights abuses – such as murder, torture, rape, and disappearances – which shouldn’t be ignored in even the most extreme situations. (Click here to read Sarah Margon’s guest post on vetting Syrian opposition groups in accordance with this law.)

Ben also says that even if the U.S. was to get involved in ground operations, its forces on the ground would be few, as would the number of detainees. This means the U.S. could, according to Ben, “leave whatever detention operations arise to local forces—keeping, perhaps a small number of high-value detainees with immediate tactical intelligence and transferring a small number of others to U.S. law enforcement for prosecution.” He concludes that a detention problem will only arise “if the U.S. ground mission grows substantially, and there too we actually have a model—one honed in the latter years in Afghanistan.”

There’s a lot to unpack here, all of which makes me believe it’s critically important to have the makings of a detention plan in place now, not later. For example, there are issues that could surface surrounding rules on potential joint interrogations, the transfer of detainees into U.S. custody, and evidence collection for law enforcement purposes.

But I want to focus on a separate issue – the principle of non-refoulement – which is also important to a U.S. detention plan.

The non-refoulement principle, as defined in the Convention Against Torture, states, “No State Party shall expel, return (“refouler”) or extradite a person to another State where there are substantial grounds for believing that he would be in danger of being subjected to torture.”

It’s likely that the U.S. won’t take the view that its non-refoulement obligations apply as a matter of law in Iraq or Syria. Nonetheless, those obligations would apply as a matter of policy (at least with respect to DoD state transfers), which is quite clear from the new DoD detention directive. (Gabor Rona raises a serious concern with the DoD non-refoulement definition here, and the  latest in a lengthy back-and-forth debate on the directive is here.)

If it turns out that U.S. ground forces do partner with groups that carry out, as Ben puts it, “abuses of captives…that [U.S.] forces would not tolerate,” then those U.S. forces, as a matter of international law and as a matter of DoD detention policy (according to DoD), would be prohibited from handing those detainees over to abusive state forces.

The question remains whether the United States would actually follow its policy on non-refoulement. In Afghanistan, it took nearly irrefutable evidence and repeated reporting from reliable organizations to push the U.S. military into ceasing transfers of its detainees to Afghan security forces – in particular the National Directorate of Security. When the U.S. military did cease detainee transfers, it determined that it could re-start the transfers only after having put in place a detailed and resource-heavy torture mitigation system (which included trainings for Afghans and detention inspections) to ensure that the abuses ended at the tainted detention facilities.

The point of this commentary is all to say that the time is not ripe to dismiss Jeff’s initial questions and concerns about U.S. detention operations when it comes to ISIS (or other groups caught up in the mix). Quite the opposite. A key lesson that should be learned from the wars in Afghanistan and Iraq is that serious detention problems arise when the U.S. minimizes the pitfalls of potential detention operations and doesn’t plan ahead. Moreover, this lesson should be applied regardless of the anticipated size of the detention population. Besides, while the number of detainees might be small at first, detention populations often balloon, and balloon quickly.