Homeland Security Secretary Markwayne Mullin stirred the pot on national TV this week when he relayed an unnamed tech contact’s theory that advanced artificial intelligence — the administration’s favored term, “SI” — could help explain a string of big outages. He also said he’s “98% certain” the incidents were not a foreign cyberattack. That claim is the new development driving headlines, and it deserves a clear look: it is newsy, it is speculative, and it points to real problems we ignore at our peril.
What Mullin said — and why it jumped into the headlines
On television, Secretary Markwayne Mullin described talking to “one guy in the tech industry” who suggested these modern AI systems might be moving too fast for old infrastructure. Mullin said the idea “kind of makes sense” and used the shorthand “SI.” He was careful to call it a theory, not a conclusion. Still, a senior DHS official bringing up a high‑stakes idea on national TV — even second‑hand — turns heads. People are rightly asking: is this a clue, a warning, or just an offhand guess that will make headlines and start rumors?
What the FAA, Microsoft, and banks are saying
The agencies and companies involved tell a different, less dramatic story so far. The FAA points to overloaded routers and a contractor upgrade as the likely cause of the nationwide ground stop and says it sees “zero evidence of anything AI‑related.” Microsoft says its outage looks like a technical issue and is doing a post‑incident review. JPMorgan Chase blamed an internal software bug for its payment delays. In short: the public, documented signs point to old gear, bad change management, and buggy code — not a wandering, self‑fixing superintelligence.
Why Mullin’s theory still matters — and what really should worry us
Whether or not “SI” had anything to do with these outages, Mullin raised a real problem: we are gluing shiny AI tools onto fragile, decades‑old systems. When banks, airlines, and government networks use legacy hardware and third‑party add‑ons, small changes can ripple into big failures. That’s the boring, avoidable threat here. Tech hype and holiday rushes to deploy new features without proper testing are far more likely culprits than a cinematic AI takeover. So yes, call out the drama, but don’t let it distract from the basic need for modernization and stronger contractor oversight.
What Washington should do next
Congress and agencies need facts, not theater. Demand FAA and contractor diagnostics, read Microsoft’s post‑mortem, get JPMorgan’s root‑cause report, and ask DHS/CISA for a technical write‑up that either backs Mullin’s claim or buries it. Praise Secretary Mullin for engaging with industry, but insist on proof before we let anonymous “one guy” theories shape policy or panic the public. Above all, use this moment to fix weak systems, tighten testing and rollback rules, and make sure critical infrastructure can handle real stress — whether from Mother Nature, human error, or a future AI feature that actually needs to be managed like any other tool.

