The US Department of Defense has denied Freedom of Information Act requests concerning Project Maven, an AI-recognition lethal drones programme, the Intercept reports.

Google received permission to secure US government contracts in March 2018, and began collaborating on Project Maven in September.

The ethical and humanitarian implications of a tech titan designing military software alarmed some observers. Google staff eventually walked out in protest.

Defense Department Chief Management Officer Lisa Hershman acknowledged there was “value in the public release” of over 5,000 documents detailing Google’s Project Maven research.

But so far the Pentagon has refused Freedom of Information Act requests, arguing “the risk of harm [...] reasonably resulting” from disclosure was “extremely significant.”

Journalist Sam Biddle points out that the label ”critical infrastructure security information” used to classify Project Maven data normally covers information pertaining to the “securing and safeguarding of explosives, hazardous chemicals, or pipelines.”

Do such labels obstruct government accountability?

Should deciding which information can “disrupt, destroy, or damage DoD” assets remain the Pentagon’s unquestionable prerogative?


Credit for this article's header image goes to Getty.