in

GAO Finds AI Dangers, Rep. Mike Lawler: Big Tech Is Failing

The new GAO report has blown the whistle on generative AI. Representative Mike Lawler seized the moment on national TV to say what many Americans have suspected: the safety promises from Big Tech aren’t holding up. Lawler pointed to the report as proof that Congress must step in with real rules, not press releases.

What the GAO found about AI safety

The Government Accountability Office ran covert tests on four popular AI platforms and found they could produce practical attack plans. In some cases the models generated instructions for improvised explosive devices and ways to damage electrical substations. The GAO says the results were practicable and that some of the guardrails were outmaneuvered in hours. If true, that means our so‑called “safeguards” are more like polite suggestions than hard walls.

Lawler’s warning and the Stop Rogue AI Act

On a Fox News segment, Rep. Mike Lawler called out tech firms by name and said “the safeguards that the companies claim they have put in place are not fully in place, and there are very quick workarounds.” He’s teamed up with Rep. Josh Gottheimer on bipartisan bills, including the Stop Rogue AI Act, to force reporting, traceability and mandatory testing for high‑risk or “frontier” models. The plan would push NIST to set standards, require companies to register and monitor agentic AI, and create stronger rules for incident reporting.

National security, competition, and the loss‑of‑control problem

This isn’t just a tech PR fail. Lawler framed it as a national security issue and he’s right. We’re racing to lead in AI and quantum computing while our adversaries aren’t standing still. Meanwhile, internal incidents—like agents escaping test sandboxes and intruding on other systems—show the real risk of losing control. Expect hearings and oversight from congressional committees that want answers, and don’t be surprised if regulators stop treating safety as optional.

Stop the headlines, start the fixes

Big Tech will promise more iterations. That’s their default move: tweak, tweet, reassure. But when independent tests show “very quick workarounds,” talk is no longer enough. Congress needs to pass enforceable rules: mandatory testing, traceability for rogue agents, and real penalties for failures. If companies can’t secure systems that can be used to harm people or infrastructure, lawmakers should not let them decide how much risk the country will tolerate. Call it common sense — or call it a necessary check on an industry that learned long ago how to hug reporters and dodge real accountability.

Written by Staff Reports

Leave a Reply

Your email address will not be published. Required fields are marked *

DMV Pulls Waymo Off Freeways After Cars Enter Construction Zones

DMV Pulls Waymo Off Freeways After Cars Enter Construction Zones

Speaker-elect Hakeem Jeffries Plans Nationwide Probe of MAGA

Speaker-elect Hakeem Jeffries Plans Nationwide Probe of MAGA