US Soldiers Are Immune from Afghan Prosecution, CRS Says
The American soldier who is accused of killing 16 Afghan civilians is under the legal jurisdiction of the U.S. government and is immune from prosecution under Afghan law, says a newly updated report from the Congressional Research Service.
Afghan officials had said they wanted the soldier to be tried in Afghanistan, not in an American military court, the New York Times and other press outlets reported.
But according to CRS, the Status of Forces Agreement (SOFA) between the US and Afghanistan dictates otherwise.
“In the case of Afghanistan, the SOFA, in force since 2003, provides that U.S. Department of Defense military and civilian personnel are to be accorded status equivalent to that of U.S. Embassy administrative and technical staff under the Vienna Convention on Diplomatic Relations of 1961,” the CRS report said.
“Accordingly, U.S. personnel are immune from criminal prosecution by Afghan authorities and are immune from civil and administrative jurisdiction except with respect to acts performed outside the course of their duties. The Government of Afghanistan has further explicitly authorized the U.S. government to exercise criminal jurisdiction over U.S. personnel.”
“Thus, under the existing SOFA, the United States would have jurisdiction over the prosecution of the servicemember who allegedly attacked the Afghan civilians.”
A copy of the CRS report was obtained by Secrecy News. See Status of Forces Agreement (SOFA): What Is It, and How Has It Been Utilized?, March 15, 2012.
This DOE Office has been achieving DOGE’s stated mission of billion dollar savings for decades. Now government leaders may close its doors.
Direct File redefined what IRS service could look like, with real-time help and data-driven improvements. Let’s apply that bar elsewhere.
At this inflection point, the choice is not between speed and safety but between ungoverned acceleration and a calculated momentum that allows our strategic AI advantage to be both sustained and secured.
Improved detection could strengthen deterrence, but only if accompanying hazards—automation bias, model hallucinations, exploitable software vulnerabilities, and the risk of eroding assured second‑strike capability—are well managed.