The standards laid out for when a targeted killing can be justified are not, in themselves, unreasonable:
First, the U.S. government has determined, after a thorough and careful review, that the individual poses an imminent threat of violent attack against the United States; second, capture is not feasible; and third, the operation would be conducted in a manner consistent with applicable law of war principles.
But the problem, as Adam says, is that a great deal of the work is done by the term “imminent threat,” and Holder’s follow-up already indicates slippage:
But don’t assume that when Holder says “imminent threat of violent attack,” he means that you’re actually part of a specific plot threatening American lives. “The Constitution does not require the president to delay action until some theoretical end stage of planning when the precise time, place, and manner of an attack become clear,” Holder said. That would introduce an “unacceptably high risk of failure.” When he refers to “failure,” Holder presumably means failing to kill the target before the attack or plan for an attack materializes, not the possibility that the government might accidentally kill an innocent person.
And it’s precisely this potential for defining “imminent threat” down that makes the lack of oversight unacceptable. If the executive branch can’t demonstrate evidence that there is an “imminent threat” to some sort of independent body, there’s no reason to believe that those being targeted for killing in fact pose imminent threats, and the potential for abusing the gravest power the executive branch possesses remains. Without meaningful oversight, these standards are only as good as the administration applying them. And that’s just not nearly good enough.