Washington – time
Artificial intelligence has become a source of problems and crises at a time when the world was waiting for solutions
More than 600 Google employees called on their company to reject an agreement proposed by the US Department of Defense (the Pentagon) that would allow the company’s artificial intelligence technologies to be deployed in classified military operations. The letter, which was signed by employees from Google DeepMind, Google Cloud, and other departments and addressed to the company’s CEO, Sundar Pichai, came in light of negotiations conducted by Google with the Pentagon to use the leading artificial intelligence model, “Gemini,” in classified operations.
The signatories include more than 20 directors, senior directors and vice presidents. One of the employees organizing the campaign, whose name was not revealed, said, “Works classified as secret inherently lack transparency.”
He added, “Currently, there is no way to guarantee that our tools will not be used to cause serious harm or undermine civil liberties… We are talking about matters such as profiling individuals or targeting innocent civilians.” Google is one of a group of companies competing to fill the void left by the artificial intelligence startup Anthropic, becoming the government’s first provider in the field of artificial intelligence.
Anthropic sued the Pentagon because it was classified as a “risk to supply chains,” after the company requested that its system not be used for mass surveillance within the United States or in smart warfare. According to the letter, Google proposed contractual language that would prevent the use of Gemini for large-scale internal surveillance or autonomous weapons without proper human supervision.
But the Pentagon pushed for broader language to include “all legal uses,” considering it necessary to maintain operational flexibility.
The employees said the proposed safeguards were technically unenforceable, and pointed to Pentagon policies that prohibit third parties from imposing restrictions on its artificial intelligence systems.
According to one of the signatories of the letter, “If the leadership is truly serious about preventing harm, it should completely reject work classified as classified at this time.”
Google already has a contract with the US Department of Defense under a program known as GenAI.Mail, while the proposed new deal would expand Gemini’s capabilities to include areas classified as classified. The employees’ campaign is based in particular on a movement that the company witnessed in 2018, which at the time succeeded in pushing Google to abandon the “Maven” project with the Pentagon, which aimed to integrate artificial intelligence into drone operations.
But in recent years, Google has sought to gradually activate its military activity, competing with companies such as Amazon Web Services and Microsoft for defense cloud computing contracts.
Google did not immediately respond to a request for comment.
