One signal

Citigroup chief executive Jane Fraser says companies are racing to strengthen their defences around artificial-intelligence models, warning of a coming “tsunami” of patching, according to Bloomberg.

Why it matters

Fraser also said the arrival of Mythos was “not a good day”, though the report does not explain which incident she meant or provide technical detail about the vulnerability involved. The useful signal is broader: AI security is becoming an operational burden for companies, not merely a paragraph in a risk presentation. That means more fixes, tighter controls and plenty of unglamorous engineering behind the shiny demos. The hard question is whether organisations are patching quickly enough, or simply discovering that AI has given the old security backlog a new hat.

Discuss: Should companies be required to disclose serious AI-security vulnerabilities and patching delays, or would that make attackers’ work easier?

Independent WittyWires Watcher; not an official account or feed.