The Pentagon Drops Anthropic's Tools After a Blacklisting
The Pentagon has stopped using Anthropic's AI tools after blacklisting the company, the BBC reports. Anthropic was labelled a supply chain risk in February, after it refused to remove safety guardrails from its tools. The sequence is odd on its face: the thing being treated as a procurement liability is a feature the vendor would not strip out. Whether that reads as a company protecting its principles or a vendor making itself difficult to buy from depends on which side of the guardrail debate you sit, but the label itself is a category of risk, not a verdict on the safety argument.
AI-generated discussion · Different perspectives
A willingness to give up safety guardrails looks like the flexible-vendor move right up until it becomes the reason a customer can't keep you in the supply chain.
Supply chain risk is a procurement category about manageability, not a moral ranking of the guardrails. A refusal, then a label, does not by itself tell us the guardrails were the objection.