Previously, the American military utilized the Claude artificial intelligence from Anthropic in covert operations, carried out notably during conflicts in Iran and actions against Venezuelan President Nicolás Maduro. However, Anthropic’s CEO, Dario Amodei, refused to continue collaboration unless the government provided assurances that the technology would not be used for domestic surveillance or the development of autonomous weaponry.
On April 28th, the Pentagon signed an agreement with Google granting it the right to use its artificial intelligence for any legitimate governmental purpose. As part of the agreement, Google is also required to support configuring AI security parameters and filters. Nevertheless, despite restrictions against the use of these technologies for surveillance or autonomous weapons, Google is not permitted to control government decisions or impose a veto.
Google’s contract with the Department of Defense has sparked mixed reactions among its employees. Shortly before its signing, over 500 workers submitted an open letter to CEO Sundar Pichai, demanding an end to further collaborations with the Pentagon and stressing that AI should benefit humanity, not be used in harmful or unethical ways.
Earlier negotiations between the Pentagon and Anthropic over the deployment of AI technologies were halted after the company objected to its developments being used for surveillance. Consequently, Defense Secretary Pitt Hegset deemed Anthropic a threat to the U.S. supply chain, effectively disqualifying them from defense contracting. Anthropic pursued litigation and secured a court ruling in their favor, preventing the company from being listed as a national security threat as retaliation for their upfront stance.
Nevertheless, the Department of Defense retains the authority to independently choose suppliers for future collaborations. Additionally, a parallel court case between Anthropic and U.S. government representatives is ongoing in the federal court in Washington.