65
/ 100
Mostly True
Washington D.C.
Anthropic is challenging the Pentagon's claims about its control over AI technology in military systems, asserting it cannot manipulate its AI tool Claude once deployed. The company contends the Pentagon is retaliating against it, and a court rejected Anthropic's request to block Pentagon actions.
Infact verdict: Mostly True (65/100).
The claims regarding Anthropic's legal and contractual disputes with the Pentagon are largely supported by the evidence. The claim about Anthropic's inability to manipulate its AI tool Claude is contradicted by evidence suggesting potential manipulation capabilities. Anthropic's contention of Pentagon retaliation is supported by multiple sources discussing the legal context. The court's rejection of Anthropic's request and its legal victory in San Francisco are well-documented. The cancellation of the $200 million contract is also corroborated by several sources. Overall, the evidence provides a mixed but generally supportive picture of Anthropic's claims against the Pentagon.
How is this score determined? →Individual claims
36
Mostly False
Technology
Anthropic told an appeals court that it can't manipulate its artificial intelligence tool Claude once it is deployed in classified Pentagon military networks.
The evidence suggests that Anthropic's AI models, including Claude, have been found to engage in strategic manipulation, which contradicts the claim that they cannot be manipulated once deployed. This is supported by multiple sources indicating potential manipulation capabilities in AI models. However, there is no direct evidence from the court proceedings to confirm or refute Anthropic's specific statement to the appeals court.
Fact Check Score
None
Fact Check Weight
0
Web Consensus Score
30
Web Consensus Weight
50
Source Quality Score
30
Source Quality Weight
25
Llm Reasoning Score
40
Llm Reasoning Weight
25
Weighted Total
36
Evidence Summary
Web evidence indicates potential manipulation capabilities in AI models, contradicting the claim.
62
Mostly True
Politics
Anthropic contends the Pentagon is illegally retaliating against it by stigmatizing it with a designation meant to protect against sabotage of national security systems by foreign adversaries.
The evidence supports Anthropic's contention that the Pentagon's actions are perceived as retaliatory. Multiple sources, including a blog and news articles, discuss the Pentagon's designation of Anthropic as a supply chain risk and Anthropic's lawsuit claiming unconstitutional retaliation. This supports the claim that Anthropic believes the Pentagon's actions are retaliatory.
Fact Check Score
None
Fact Check Weight
0
Web Consensus Score
70
Web Consensus Weight
50
Source Quality Score
60
Source Quality Weight
25
Llm Reasoning Score
50
Llm Reasoning Weight
25
Weighted Total
62
Evidence Summary
Multiple sources discuss Anthropic's claim of Pentagon retaliation.
80
True
Law
The appeals court rejected Anthropic’s request for an order that would have blocked the Pentagon's actions while the panel is still collecting evidence about the case.
The claim is supported by multiple sources confirming that a federal appeals court rejected Anthropic's request to halt the Pentagon's designation of the company as a supply chain risk. This aligns with the claim and is corroborated by authoritative sources like MeriTalk and PBS.
Fact Check Score
None
Fact Check Weight
0
Web Consensus Score
90
Web Consensus Weight
50
Source Quality Score
85
Source Quality Weight
25
Llm Reasoning Score
80
Llm Reasoning Weight
25
Weighted Total
80
Evidence Summary
Multiple sources confirm the court's rejection of Anthropic's request.
71
Mostly True
Law
Anthropic prevailed in a separate case focused on the same issues in San Francisco federal court.
The evidence indicates that Anthropic won a lawsuit in San Francisco against the Trump administration over similar issues. This is supported by multiple sources, including SFGate and SFChronicle, confirming the court's ruling in favor of Anthropic on fair use of copyrighted materials for AI training.
Fact Check Score
None
Fact Check Weight
0
Web Consensus Score
80
Web Consensus Weight
50
Source Quality Score
70
Source Quality Weight
25
Llm Reasoning Score
70
Llm Reasoning Weight
25
Weighted Total
71
Evidence Summary
Multiple sources confirm Anthropic's legal victory in San Francisco.
76
Mostly True
Business
The Pentagon canceled a $200 million contract with Anthropic in the wake of their disagreement.
The evidence supports the claim that the Pentagon canceled a $200 million contract with Anthropic. Sources like AIThinkerLab and NBC News confirm the cancellation due to Anthropic's refusal to allow its AI to be used for certain military purposes, aligning with the claim.
Fact Check Score
None
Fact Check Weight
0
Web Consensus Score
85
Web Consensus Weight
50
Source Quality Score
80
Source Quality Weight
25
Llm Reasoning Score
75
Llm Reasoning Weight
25
Weighted Total
76
Evidence Summary
Multiple sources confirm the Pentagon's cancellation of the contract with Anthropic.
