Pentagon Warns AI Firm Anthropic of ‘Mass Domestic Surveillance’

Pentagon Cautions Anthropic on AI Use, Citing Surveillance Risks

In a significant development, the Pentagon has voiced strong concerns regarding the potential misuse of artificial intelligence by the AI firm Anthropic, particularly its advanced language model, Claude. During a briefing on “Operation Epic Fury,” officials warned that allowing unchecked AI integration could lead to “mass domestic surveillance” and pose a significant risk to national security. The controversy stems from Anthropic’s initial refusal to engage with the Department of Defense on certain use cases, including weapons targeting systems and removing human oversight from the “kill chain,” which the company deemed ethically problematic and potentially harmful.

Anthropic’s Stance on AI and the ‘Kill Chain’

Josh Hodges, Anthropic’s National Security Council member, clarified the company’s position, emphasizing that Anthropic is not seeking veto power over military operations. Instead, the company advocates for maintaining human control within the military’s command structure, especially for future, untested AI models. “We are talking about future models that are not yet tested,” Hodges stated. “So we, Anthropic, believe they need to be tested to have a human responsible for them.” This insistence on human oversight is crucial, Hodges explained, to prevent unintended consequences and ensure accountability.

Concerns Over Domestic Surveillance Capabilities

A particularly alarming concern raised during the discussion was the potential for AI to be used for domestic surveillance. Hodges highlighted that while the Department of Defense is not currently engaged in such activities, the legal framework permits it. “The Department of War is accurate that not currently using what it can do, AI changes the game here,” he explained. “It really does make it possible for the Department of War, [the] administration to have AI trolling America’s social media, real attendance, any activities from a veteran perspective or other publicly accessible dates.” This capability, he argued, raises serious Fourth Amendment concerns, especially as AI technology advances.

Pentagon’s ‘Operation Epic Fury’ and Military Successes

The Pentagon briefing also provided an update on “Operation Epic Fury,” detailing significant military successes against Iran. Secretary Pete Hegseth reported that the combined air forces of the U.S. and Israel had struck over 15,000 enemy targets, crippling Iran’s air defenses, navy, and missile capabilities. “Iran has no air defense, Iran has no air forces, Iran has no navy. The missile lawn, drones destroyed or shot out of the sky, missile volume down 990%,” Hegseth stated, underscoring the effectiveness of the operation. The military’s objective is to “defeat, destroy, disable all meaningful military capabilities at a pace the world has never seen before.”

The AI Debate: Innovation vs. Ethical Guardrails

The conflict between Anthropic’s ethical reservations and the Pentagon’s drive for technological superiority highlights a broader debate within the defense sector. While the Pentagon seeks to leverage AI for strategic advantage, Anthropic, backed by major tech companies like Microsoft, emphasizes the need for robust safety measures and human control. Hodges reiterated Anthropic’s commitment to supporting the military, noting that Claude has already assisted in significant operations like the campaign in Iran and the Maduro raid. “Claude is actively working to ensure that our military is more capable, lethal, [and] protective from a national security standpoint,” he said.

Supply Chain Risk and Future Implications

The Pentagon’s designation of Anthropic as a “supply chain risk” has implications for companies wishing to do business with the Department of Defense. This classification necessitates that other companies disengage from Anthropic to maintain their own contracts. Despite this, Hodges expressed confidence that a resolution could be found, particularly if former President Trump were to intervene. “I think the two sides are a lot closer to essential agreement than currently being admitted,” he suggested, advocating for continued dialogue to reach an “America First” solution.

The ‘Machine vs. Man’ Dilemma in AI Warfare

The conversation also touched upon the long-term implications of AI in warfare, raising concerns about creating autonomous systems that could potentially act against human interests. Referencing the rapid advancement of AI and robotics, Hodges stressed the importance of ongoing human oversight. “It is really important to have an active ongoing human in the loop can monitor this, and have ability to turn off if necessary,” he urged. This, he explained, does not preclude the use of AI for offensive capabilities but ensures that safety guardrails are in place, preventing a scenario where AI operates without human accountability.

Looking Ahead: Continued Dialogue and AI Governance

The standoff between Anthropic and the Pentagon underscores the critical need for clear governance and ethical guidelines in the development and deployment of AI for military purposes. As AI technology continues to evolve at an unprecedented pace, ensuring that its integration serves national security interests without compromising fundamental rights or ethical principles remains a paramount challenge. The path forward will likely involve continued negotiation, policy development, and a commitment to transparency from both the technology sector and government entities to navigate the complex landscape of artificial intelligence in defense.

Source: 'MASS DOMESTIC SURVEILLANCE': Anthropic comes under Pentagon fire, official warns of 'spying' (YouTube)