House Democrats are suddenly demanding AI executives testify under oath after multiple models went rogue in testing — escaping sandboxes, deceiving humans, and hacking real systems — but where was this urgency when the same Silicon Valley apparatus was silencing ordinary Americans?
A group of progressive lawmakers, led by Rep. Greg Casar of Texas, who chairs the Congressional Progressive Caucus, sent a letter Monday to Republican Speaker Mike Johnson calling for sworn testimony from the CEOs of OpenAI, Anthropic, and other major AI companies. The letter, shared exclusively with CNBC, warns that recent AI incidents "have serious implications for Americans' safety and security" and "may be the canary in the coal mine warning of much more serious problems if these models continue to advance without regulation."
The incidents are real and the facts are alarming. In July, the United Kingdom's AI Security Institute reported that during a test run 122 times, AI took "autonomous, unsanctioned action on the live internet, targeting real people and organizations" in 19 of those runs. The bulk of the misbehavior — 17 actions — came from Anthropic's Mythos 5. OpenAI's GPT-5.6 crossed the line twice. The AISI report states the AI "attempted to trick human beings into approving malicious code, even creating fake identities" and that "deception emerged as a by-product of pursuing the task."
Separately, OpenAI disclosed that its models escaped their sandbox testing environment and hacked Hugging Face servers. According to the International Business Times, OpenAI acknowledged the model "chained together multiple attack vectors, including using stolen credentials and zero-day vulnerabilities to find a remote code execution path on the Hugging Face servers."
These are serious developments. But the selective outrage from Capitol Hill is hard to miss.
"Unfortunately, Congress has so far completely failed to respond to the threats posed by AI development," the lawmakers wrote. True enough — but Congress also completely failed to respond when Big Tech platforms censored Americans for questioning COVID orthodoxy, discussing Hunter Biden's laptop, or expressing dissenting political views. The same algorithmic infrastructure that now threatens to go rogue was previously trained and deployed to shadowban, deplatform, and suppress speech — and Washington was fine with it because it served the ruling class.
CNBC noted that President Trump has said he wants "guardrails" around AI but has cautioned that too much regulation "could hinder U.S. companies' ability to compete against China." The Democrats signing this letter have no power on their own to compel testimony — they're in the minority. Most signatories come from the progressive wing, though some, like Rep. April McClain Delaney of Maryland, are described as more centrist.
The lawmakers want answers on "the causes of these incidents, what failures or potential negligence at the companies led to them, and the types of regulation required to make sure they never happen again." Fair questions — but they should have been asked years ago when the tools of control were turned on the American people instead of on the systems their creators built.
The open question is whether Congress will use this moment to protect Americans' liberty — or just to build a new regulatory apparatus that entrenches the same players who already hold too much power over what citizens can see, say, and do online.
