iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Jamie Dimon reportedly said AI risks “went up 10-fold after Mythos,” but that is his assessment—not a published measurement of a tenfold rise in cyberattacks, losses, or successful breaches. The report quoting him provides no baseline or calculation. Anthropic has separately described Mythos’s vulnerability-finding capabilities and reported findings from defensive programs, but those figures do not verify Dimon’s multiplier.
What did Jamie Dimon say about Mythos and cyber risk?
In a Bloomberg TV interview on October 6, 2026, JPMorgan Chase CEO Jamie Dimon reportedly said risks from AI “went up 10-fold after Mythos.” The Business Times report, published October 7, also quoted him saying AI had created vulnerabilities “that we didn’t know about” and calling the downside from agents and Mythos a legitimate concern. He said JPMorgan was working to address the problem. The Business Times report is the source for these remarks.
The report does not define “risk” or explain what was compared before and after Mythos. It supplies no underlying data, calculation, or independent validation. The figure should therefore be read as Dimon’s reported risk assessment—not as a finding that attacks, vulnerabilities, or financial losses increased tenfold.
What is Mythos, and why has it raised concern?
Anthropic describes Claude Mythos Preview as an unreleased general-purpose model with advanced coding capabilities that can identify and exploit software vulnerabilities. The company says it found thousands of high-severity flaws across major operating systems and browsers, and introduced Project Glasswing to apply such capabilities to defending critical software. Those are Anthropic’s claims about its model and initiative, not an independent measurement of real-world cyber risk. Anthropic’s Project Glasswing announcement provides its account.
#1 Best Overall
The concern is dual-use: finding a flaw can help a defender fix it, while knowledge of the same flaw could help an attacker exploit it. A model’s ability to discover vulnerabilities is not the same thing as evidence that attackers have used it successfully, or that a particular number of incidents or losses has followed.
What do Anthropic’s vulnerability figures show?
Anthropic has reported several totals from different programs and time periods. They count vulnerabilities found, not cyberattacks or losses, and should not be combined as though they were a single measure of threat growth.
| Reported figure | What it covers | Important qualification |
|---|---|---|
| More than 10,000 high- or critical-severity vulnerabilities | Initial findings by roughly 50 Project Glasswing partners, reported in May 2026 | Anthropic’s account of partner findings during the effort’s initial weeks. Source: Anthropic’s May update. |
| At least 129,000 verified vulnerabilities | Project Glasswing partner findings from April through July 2026 | Anthropic said more than 33,000 were rated high or critical. It cautioned that the total likely undercounted findings because it relied on survey data from only some partners. Source: Anthropic’s Cyber Verification Program announcement. |
| 5,500 verified vulnerabilities | Anthropic’s own open-source scanning from April through October 2026 | A separate company-reported program and period; it is not the same scope as the partner total. Source: Anthropic’s Cyber Verification Program announcement. |
These numbers indicate that Anthropic and its partners say they have identified many software flaws through defensive work. They do not tell readers how many flaws were exploited, how many incidents occurred, or whether overall cyber risk rose by a particular factor.
Free tools Windows power users keep installed
One-click scans. No signup required.
Who can use Mythos?
Mythos is not generally available to the public. Anthropic’s current model page says Mythos 5.1 is limited to vetted organizations through trusted access programs. Anthropic’s Mythos page describes that access status.
Rank #3
Anthropic’s Cyber Verification Program announcement describes three access tiers for qualifying security teams, with advanced capabilities and reduced blocking classifiers tied to the scope of cyber work. The company says generally available models retain conservative cyber safeguards and that it will continue refining tier-based classifiers. These controls describe Anthropic’s approach; they do not independently establish how much risk the model creates or prevents.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What would establish whether cyber risk actually rose?
A tenfold estimate needs a defined measure and a comparable baseline. For example, a change in the number of vulnerabilities discovered is not interchangeable with a change in successful intrusions, expected financial losses, or the exposure of a particular organization. Dimon’s reported comment does not specify which meaning he intended.
To assess the real-world effect, readers would need evidence that connects model capability and access to outcomes: verified exploitation, incident trends, losses, or changes in exposure measured over a defined period. The available reporting gives no such calculation for Dimon’s multiplier. Anthropic’s defensive-program totals offer relevant context about vulnerability discovery, but they cannot supply that missing measure.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
