President Trump issued a memorandum last week that transfers to the Attorney General the authority of the Director of National Intelligence to declassify intelligence information concerning the 2016 election.
The memorandum effectively amends Executive Order 13526 on classification on national security information, but in a highly customized way: It applies only to Attorney General William Barr (not any successors) and only to the investigation of the 2016 presidential campaigns. The memorandum was published in the Federal Register today.
Even so, the move represents a functional demotion of the Director of National Intelligence and a partial transfer of his authority to the Attorney General.
Executive Order 13526 gave sweeping authority over declassification of intelligence information to the DNI, who was authorized to “declassify, downgrade, or direct the declassification or downgrading of information or intelligence relating to intelligence sources, methods, or activities.” (sect. 3.1c)
The new presidential memorandum adopts the same language but modifies the provision to state that it is Attorney General Barr who may now “declassify, downgrade, or direct the declassification or downgrading of information or intelligence that relates to the Attorney General’s review.”
No rationale for the change was provided, though it was understood to support the Attorney General’s investigation into what he called U.S. government “spying” on the Trump campaign.
Senator Mark Warner (D-VA) warned that the move threatened to politicize intelligence. “Selectively declassifying sources and methods in order to serve a political agenda will make it harder for the intelligence community to do their jobs protecting this country from those who wish to do us harm,” he said.
For his part, DNI Dan Coats said that “I am confident that the Attorney General will work with the IC in accordance with the long-established standards to protect highly-sensitive classified information that, if publicly released, would put our national security at risk.”
* * *
There is some precedent for overriding the judgment of the DNI concerning the protection of sources and methods.
A 1999 decision of the Justice Department Office of Legal Counsel concluded that the presidentially-established Interagency Security Classification Appeals Panel could declassify intelligence information over the objections of the Director of Central Intelligence.
While it is true that the DCI, and now the DNI, is obliged by the National Security Act to protect intelligence sources and methods from unauthorized disclosure, the Director’s authority in this area is not absolute or exclusive.
Specifically, “If the President concludes that information concerning intelligence sources and methods should not be classified, the disclosure of such information simply is not ‘unauthorized’ within the meaning of the [National Security Act],” wrote Randolph D. Moss of the Office of Legal Counsel in his 1999 opinion.
Still, this OLC conclusion may not be correct (said a non-lawyer) because “declassification” is not the same as “disclosure.” Even intelligence information that is declassified or unclassified may still be, and often is, protected from public disclosure by the DNI under the provisions of the National Security Act.
The new presidential memorandum does not address the question of disclosure at all.
Commercial artificial intelligence tools have recently emerged that are able to produce police reports. If the resulting reports are inaccurate, incomplete or biased, or if the process leaks confidential information, this could undermine the criminal justice system and harm citizens.
Too often, affected patients, clinicians, and regulators cannot see how the system works, why a decision was made, or whether meaningful human oversight occurred.
Existing tools from other domains, such as existing robust public engagement processes in drug development, when applied to AI deployment can help strengthen public trust in these systems and enhance perceptions of their legitimacy and the decisions they produce.
With thoughtful policy action, it is still possible to build systems that are fair, transparent, and accountable, and to earn the public trust that will ultimately determine AI’s future. We hope policymakers are ready to act.