Mike Johnson is letting AI monsters escape for corporate profit
By Imani Sutton · Reporting from Atlanta ·
The CEOs of Anthropic, OpenAI, and xAI have finally admitted the truth: they’ve built a car with no brakes, and they’re terrified of the cliff.
The CEOs of Anthropic, OpenAI, and xAI have finally admitted the truth: they’ve built a car with no brakes, and they’re terrified of the cliff. According to CNBC, these industry leaders have pledged to slow development, while a former researcher, Jacob Coxon, told ABC News that the people building this tech "earnestly believe that it could kill us all by the end of the decade."
When the people making the product tell you it’s lethal, you don't ask for a "conversation." You call the fire department. But in Washington, the fire department is currently arguing over who gets to hold the hose while the building burns. Senator Ruben Gallego hit the nail on the head, noting that "Dr. Frankenstein is telling us the monster is escaping," yet the legislative response is a shrug and a calendar invite for after the midterms. We’ve seen this movie before. During the Financial Crisis of 2008, we watched the government scramble to regulate a complex system only after the tipping point had already been crossed and the houses were already gone.
The China Shield is a Corporate Cloak
The strongest argument against urgent regulation is the "China Race." Speaker Mike Johnson claims that if Congress rushes into emergency sessions, "we will lose the race to China," a sentiment echoed by President Donald Trump, who insists that "whoever wins AI wins." This is the ultimate trump card for the deregulation crowd: the idea that safety is a luxury we can't afford if we want to maintain global hegemony.
But that logic is a ghost story designed to keep us compliant. As Senator Gallego correctly pointed out, "AI does not care once it becomes fully dangerous whether we’re Chinese or American." A systemic collapse or an agentic AI catastrophe doesn't stop at a border. The "innovation" argument is just a shield for the companies currently enjoying valuations in the stratosphere—Anthropic at $965 billion and OpenAI at $852 billion. They aren't worried about China; they're worried about their margins. We saw this same legislative deadlock during the 2021 cryptocurrency crash, where "disruption" was used as a synonym for "immunity from the law" until the bubble burst.
We Need a Sentry, Not a Committee
Right now, the "solutions" on the table are laughable. The FRONTIER Act, led by Reps. Jay Obernolte and Lori Trahan, suggests "kill switch" requirements and antitrust waivers. A "kill switch" for a distributed, frontier model is a press-release fantasy; it’s like trying to put a leash on a weather pattern. The Senate, where John Thune, Amy Klobuchar, and Ted Cruz are drafting a bill, is moving with the urgency of a glacier.
If we actually wanted to manage systemic risk, we would look at the Pakistan Environmental Protection Agency establishment. The Pak-EPA wasn't a series of "guardrails" or a bipartisan committee; it was the creation of a specialized federal agency tasked with protecting human health and the environment through actual enforcement mechanisms. AI is not a "tool"; it is a systemic industrial externality. It changes the load on our labor markets, our energy grids, and our cognitive security. You don't manage that with a "big room" and a handshake, as Speaker Johnson suggests. You manage it with a dedicated agency that has the power to pull the plug without asking the CEO for permission. We tried the "wait and see" approach with the Cambridge Analytica scandal, and we ended up with a digital landscape where the algorithm decides the truth.
The Midterm Nap
The most galling part of this scramble is the timing. While Democrats led by Rep. Sam Liccardo are writing letters warning that our inaction will be "inexplicable and unforgivable" to a post-apocalyptic history, Speaker Mike Johnson is preparing to send the House home until after the November midterms.
This is the same frantic, lagging energy we saw during the COVID-19 pandemic, where we deployed technology under emergency pressure while regulators tried to write the safety standards in real-time. Only this time, the risk isn't a virus; it's a decision-making engine that operates faster than the law.
The House will not pass meaningful guardrails before November. The Senate will likely produce a watered-down compromise that protects the "innovation" of the few at the expense of the many. We are choosing to let private entities steer the species toward a cliff because we're too afraid of losing a race that the winners won't even survive. Washington isn't scrambling to save us; it's scrambling to make sure the corporations are still profitable when the monster finally gets out.