# Anthropic’s AI model sparks rush from industry, government to batten down defence hatches.

*fintech · news · 2026-04-13 · The Globe and Mail*

## Key points

- Anthropic's new AI model, Mythos, can rapidly identify unknown vulnerabilities in legacy software systems.
- Mythos was tested exclusively with about 50 U.S. companies, excluding Canadian banks and platforms.
- Early Mythos tests uncovered 'zero day' flaws in systems over a decade old, raising security concerns.
- Anthropic accidentally leaked Mythos descriptions and publicly posted Claude Code's source code in March 2024.
- Canadian financial regulators and banks are demanding equal early access to Mythos for defensive readiness.

**Companies:** Amazon, Apple, Microsoft, JP Morgan Chase, Bank of America, Royal Bank of Canada, Toronto-Dominion Bank, CIBC, Bank of Montreal
**Countries:** Canada, United States

[Read the full story on The Globe and Mail](https://www.theglobeandmail.com/business/commentary/article-anthropic-mythos-ai-glasswing-hackers-defence/)

---

Canonical: https://newsio.io/n/10f43ea7-5932-4858-b688-b44209f78f8d/anthropics-ai-model-sparks-rush-from-industry-government-to-batten-down-defence
Summarized by Newsio from The Globe and Mail. https://newsio.io/how-it-works