The Pentagon has formalized agreements with seven major technology companies to integrate their artificial intelligence (AI) capabilities into its classified computer networks. This strategic move aims to enhance military operations by leveraging AI to augment warfighter decision-making in complex operational environments. The companies involved in these groundbreaking deals include Google, Microsoft, Amazon Web Services, Nvidia, OpenAI, Reflection, and SpaceX.

Notably absent from this list is AI company Anthropic, which previously engaged in a public dispute and legal battle with the Trump administration over the ethical and safety implications of AI usage in warfare. This absence underscores the ongoing debate surrounding the responsible deployment of advanced AI in military contexts.

The Defense Department has been rapidly accelerating its adoption of AI in recent years, recognizing its potential to significantly reduce the time required to identify and strike targets on the battlefield. Furthermore, AI can streamline critical logistical processes, such as organizing weapons maintenance and optimizing supply lines, as highlighted in a March report from the Brennan Center for Justice.

However, the military's increasing reliance on AI has raised substantial concerns. These include the potential for AI tools to invade Americans' privacy and, more critically, the possibility of machines autonomously choosing targets on the battlefield. Such fears gained prominence during Israel’s conflict against militants in Gaza and Lebanon, where U.S. tech giants were reportedly assisting Israel in tracking targets, coincident with a surge in civilian casualties. This fueled anxieties that these powerful tools might contribute to the deaths of innocent people.

Helen Toner, interim executive director at Georgetown University’s Center for Security and Emerging Technology and a former OpenAI board member, emphasized the role of AI in modern warfare. She noted that many command center decisions in fast-moving, confusing situations could benefit from AI systems capable of summarizing information or identifying potential targets from surveillance feeds. Yet, Toner stressed that crucial questions regarding appropriate levels of human involvement, risk assessment, and operator training are still being addressed. The challenge lies in rapidly deploying these tools for strategic advantage while ensuring operators are adequately trained and avoid automation bias, a phenomenon where individuals tend to over-trust machines.

Anthropic's dispute with the Pentagon stemmed from its demand for contractual assurances that its technology would not be used in fully autonomous weapons or for the surveillance of Americans. Defense Secretary Pete Hegseth, however, insisted that the company must permit any uses the Pentagon deemed lawful. The situation escalated when President Donald Trump attempted to prohibit federal agencies from using Anthropic’s chatbot, Claude, and Hegseth sought to label the company a supply chain risk, a designation typically used to safeguard national security systems from foreign sabotage.

In Anthropic's stead, OpenAI announced a deal with the Pentagon in March to provide ChatGPT for use in classified environments, an agreement it reconfirmed on Friday. OpenAI stated its belief that those defending the United States should have access to the best tools available. One company's agreement with the Pentagon specifically stipulated the necessity of human oversight for any missions involving autonomous or semi-autonomous AI systems. It also mandated that AI tools be used in a manner consistent with constitutional rights and civil liberties, echoing Anthropic's earlier concerns, though OpenAI claims to have secured similar assurances.

Emil Michael, the Pentagon’s chief technology officer, explained the strategic rationale behind engaging multiple providers to CNBC, acknowledging the friction with Anthropic. He stated that it would have been irresponsible to depend on a single company, especially after learning that one partner was unwilling to collaborate on the Pentagon's terms. Consequently, they sought out diverse providers.

While companies like Amazon and Microsoft have a long history of working with the military in classified settings, others, such as chipmaker Nvidia and the startup Reflection, are new to such partnerships. Both Nvidia and Reflection develop open-source AI models, which Michael regards as a priority to establish an “American alternative” to China's rapid advancements in AI systems, many of which have publicly accessible components.

The Pentagon affirmed on Friday that military personnel are already utilizing its AI capabilities through its official platform, GenAI.mil. The department stated that “Warfighters, civilians and contractors are putting these capabilities to practical use right now, cutting many tasks from months to days,” emphasizing that these growing AI capabilities will empower warfighters to act confidently and safeguard the nation against threats. AI applications range from predicting helicopter maintenance needs and optimizing troop and gear logistics to distinguishing between civilian and military vehicles in drone surveillance feeds. However, the risk of automation bias, where humans assume machines perform better than they actually do, necessitates careful implementation and operator training.