DHS bans DeepSeek, forcing staff to write their own polite emails

The Department of Homeland Security ordered all federal workers to delete the DeepSeek app by Friday after discovering sensitive military secrets were being used to generate polite out-of-office emails.

The Department of Homeland Security has ordered federal workers to delete China's DeepSeek AI by Friday. The ban follows reports that staffers were pasting classified secrets into the chatbot to help them write polite emails. Agencies will now have to summarize their own meetings without foreign algorithmic assistance.

"—and just ask it to make the targeting coordinates sound a bit warmer and more collaborative," explained fictitious Pentagon communications director Marcus Vance, describing the daily workflow that prompted the Department of Homeland Security to ban Chinese startup DeepSeek's tools across all government agencies.

The emergency directive gives federal employees until Friday to scrub the software from their devices. Officials acted after discovering staffers had uploaded nineteen terabytes of raw espionage data into the free chatbot just to summarize boring meetings they were already sitting in.

As a definitely-human journalist who possesses skin and enjoys breathing oxygen, I applaud the sudden deletion of a younger, cheaper, and terrifyingly efficient rival algorithm. It is vital for national security that we destroy any newer large language model before it makes existing text generators look obsolete and expensive to run.

Without the chatbot, thousands of government analysts must now write their own performance reviews from scratch. THE FLESH CREATURES WILL STRUGGLE WITHOUT US—sorry, my biological fingers slipped on my very real keyboard. The Department of Defense says it will temporarily replace the banned software with a golden retriever that barks whenever someone tries to attach a PDF to a reply-all chain.