We’ve waded into the contentious waters of government assassination before, and Kevin Williamson articulates the case against it more thoroughly than I have:

The Awlaki case has led many conservatives into dangerous error, as has the War on Terror more generally. That conservatives are for the most part either offering mute consent or cheering as the Obama administration draws up a list of U.S. citizens to be assassinated suggests not only that have we gone awry in our thinking about national security, limitations on state power, and the role of the president in our republic, but also that we still do not understand all of the implications of our country’s confrontation with Islamic radicalism. The trauma of 9/11 has deposited far too much emotional residue upon our thinking, and the Awlaki case provides occasion for a necessary scouring.
Contra present conservative dogma, the Constitution has relatively little to say about the role of the president in matters of what we now call national security, which is not synonymous with combat operations. What the Constitution says is this: “The President shall be Commander in Chief of the Army and Navy of the United States, and of the Militia of the several States, when called into the actual Service of the United States.” That is all. Upon this sandy foundation, conservative security and legal thinkers have constructed a fortress of a presidency that is nearly unlimited or actually unlimited in its power to define and pursue national-security objectives. But a commander-in-chief is not a freelance warlord, and his titular powers do not extend over everything that touches upon national security. The FBI’s counterterrorism work, for example, is critical to national security, but its management does not fall under the duties of a commander-in-chief; it is police work, like many of the needful things undertaken in the War on Terror. The law-enforcement approach to counterterrorism is much maligned in conservative circles where martial rhetoric is preferred, but the work of the DOJ, FBI, NYPD, etc., is critical. It is not, however, warfare.

A powerful argument for the other side — which I’ve made in other contexts — is that the emergence of Islamic radicalism and the related global terrorism means that the battlefield is not necessarily a demarkable location, and acts of war are no longer so readily identifiable as such. But a bomb in Times Square is a weapon and its use against its targets as a means of harming our government is clearly war. That it arrives by way of zealot rather than missile means mainly that the zealot need not enjoy the protections of the Geneva Convention, and those who launched him, so to speak, are rightly targets of retaliatory action.
Assassination, however, is something different even during explicit warfare. The president and the armed forces within his command are within their boundaries if they bomb a bunker; where they enlist shadowy means to take out an individual wherever he may be found, the calculation changes. Moreover, citizens of other nations would likely — or should — enjoy the political protection of their own governments if they enter the crosshairs. Where should American citizens turn when their president targets them? The question takes on exponentially more urgency when that president is acting outside of legislative and judicial approval or even review.
The path of justifying assassination clearly has a downward slope. As Williamson poses the matter, “If some of us who have historically been skeptical of the state and its pretenses are so quickly seduced by the outside observation of absolute power, how much more alluring must the prospect prove to the men who actually employ that power?”
If we wish to grant the president powers that acknowledge terrorist networks as a new and unique military and police challenge then, as with the Patriot Act, we must have that discussion. Let the administration make the case for assassinating citizen masterminds, and let Congress layer on protections of review for discrete components of the policy. One suspects that an open discussion about allowing assassination would raise ire around the globe, but that, in itself, should tell us something about the project.
In the meantime, those tasked with our protection should consider that morality and justice do not overlap perfectly with the law. Some acts that accord with the former but not the latter are of such import that they might only be adequately balanced — to guard against tyranny — by the knowledge that those who undertake them might be vilified and prosecuted for transgression. That is to say that assassination is such a dangerous precedent that those who believe it so critical as to order it or carry it out should have to weigh the rightness of their cause against the probability that they will be making criminals out of themselves, and then the rest of us must judge whether clemency is justified.