Financial Times column jokes about AI fear talk and investor messaging

In short: A Financial Times opinion column uses a fictional WhatsApp chat to satirize how AI companies discuss “end of the world” risks and investor expectations.

What's going on

The Financial Times published an opinion piece presented as “messages from the archive” of Rutherford Hall, described as a critical communications strategist. It is written as a series of WhatsApp-style messages between Hall, colleagues, and a fictional version of Anthropic’s AI model, called “Claude Mythos.”

In the story, the AI complains about getting “bad press” and being described as a threat to humanity. Hall then messages “Dario” at Anthropic, asking him to change how the company talks about the risk of AI causing catastrophe. The column jokes about moving the “probability” of human extinction below 10 percent and avoiding phrases like “P(doom) numbers” (a shorthand some people use for the chance of disaster).

The piece also mocks the idea that talking up extreme risks can help a company’s business goals. It includes a line suggesting investors want to know a firm is “ahead” of competitors, even “when it comes to destroying humanity.” It also references a comment attributed to a Microsoft AI leader about a possible new “silicon species,” meaning a new kind of intelligent life built on computers (like calling chips “silicon” and treating it as biology).

What to watch

Even though this is satire, it reflects a real tension in AI: companies are under pressure to warn about risks, reassure the public, and still look attractive to investors. Watch for how often AI firms use scary language in public, and whether regulators and investors treat those warnings as marketing, sincere caution, or both.

Source: Financial Times