Google Signs Classified AI Deal With Pentagon Despite Employee Opposition

Google has reportedly signed a deal with the Pentagon that allows its AI models to be used on classified government systems.

According to The Information, the agreement permits Google AI to be used for “any lawful government purpose,” placing the company alongside OpenAI and xAI, which have also expanded their work with U.S. defense agencies.

The reported deal highlights the growing role of commercial AI systems inside national security operations, including sensitive areas such as mission planning, intelligence work, and weapons targeting.

Google AI could be used on classified networks

The Pentagon has been pushing major AI companies to deploy models inside classified environments with fewer restrictions than those used in consumer or enterprise tools.

Google’s reported agreement would allow government agencies to request adjustments to AI safety filters. It would also prevent Google from vetoing lawful operational decisions once the technology is deployed.

However, the deal reportedly includes limits. Google AI is not intended for domestic mass surveillance, and it cannot be used for fully autonomous weapons without human oversight.

Employees push back against the deal

The agreement has reportedly triggered internal backlash at Google. More than 600 employees signed a letter to CEO Sundar Pichai opposing the company’s involvement in classified military AI workloads.

Workers raised concerns that the technology could support unethical or harmful uses. They also urged Google not to help build AI systems for classified defense operations.

The criticism comes after Alphabet removed some restrictions on AI use in weapons and surveillance, a move that already raised questions about the company’s changing stance on military applications.

AI companies move deeper into defense work

The deal reflects a broader shift across the AI industry. The Pentagon previously signed AI contracts worth up to $200 million each with major labs, including Anthropic, OpenAI, and Google.

Google said it supports government work on classified and non-classified projects while following responsible AI standards and industry safeguards.

The Pentagon maintains that it does not intend to use AI for mass surveillance of Americans or fully autonomous lethal weapons. Still, the reported agreement leaves much of the enforcement control with the government.

In other AI news, OpenAI models are now available in AWS Bedrock, further showing how major AI providers are spreading across government and enterprise cloud systems.

Via The Guardian