Most company filings bury their worst fears in small print. Anthropic's runs to 92 pages. When the maker of the Claude chatbot filed to list on the US stock market on 26 October, the risk factor section of its filing became, in effect, a long and unusually candid self-portrait of an industry that is selling a technology while also worrying about it. The document does two things at once.

It lists the factors that could damage Anthropic's business, as every such filing must. But it also acknowledges the dangers that artificial intelligence poses to society and to human beings. In most sectors, those would be separate conversations: one for regulators and ethicists, another for shareholders. Here they share the same pages, and investors will be asked to price them together.

That is why the listing matters beyond one company. Anthropic is seeking what could be a record-breaking valuation, and reporters such as Johana Bhuiyan of the Guardian have described the offering as a key test for the AI boom. Private funding rounds rely on the confidence of a small circle of backers. A public listing exposes a company to thousands of strangers, quarterly scrutiny and the sharp judgment of markets that can turn quickly.