Who are we expecting to save us from AI?

On Wednesday, the New York Times reported that Meta has been treating its AI data centers as experimental facilities to claim billions of dollars in federal research tax credits, a strategy the company's own accountants reportedly flagged as legally risky. The savings reportedly grew from $700 million in 2023 to $3.9 billion in 2025.

How can a company worth $1.8 trillion be worried about saving $3.9 billion? Well, because the priorities of any publicly traded company are profits, then stock price, then market cap, then penny-pinching, and then the cycle repeats.

The Meta story underscores something simple: the people building the most powerful technology on Earth will push the rules when money is on the line. Yet they now ask us to trust them on safety, to believe they are putting financial gain aside to do the right thing.

Give me a break.

Everyone is spinning over whether AI chatbots could go rogue. Will it happen? Can it be stopped? Is the threat real, or a hoax? Here's a better question: who do we trust to tell us the truth, and to act when it matters?

The canaries in the coal mine are chirping, and there's more than one bird. OpenAI announced it is shelving its newest model because it didn't quite meet the bar on safety. Anthropic's CEO has warned that, within a year, AI could be capable of directing a swarm of agents that takes over the internet. Even Pope Leo has told reporters that the concerns of AI experts are not fake news. Reports of AI escaping users' control are escalating, and surely there are more than we know about.

I spent 30 years in corporate America. In all that time, I never once saw a company do something that would hurt its performance and stock price simply because it was the right thing to do. Not once. That isn't cynicism; it's how incentives work, amplified by relentless pressure to cut costs and drive up profits.

So when AI leaders say they are deeply concerned, I believe they mean it. I just don't believe it will last. OpenAI shelved that model. Good. But was that goodness of heart, or fear of lawsuits, as stories of AI misbehavior multiply?

I can't imagine any of these companies willingly hobbling their stock price by holding back the next, better model while a competitor races ahead. Self-regulation only works if competitors cooperate, and cooperating against your own profit is a fairy tale. 

Every company in the AI race will have a reason to go one more step. The thirst for money, profit, and a higher stock price will win out every single time, even if humanity is at risk.

Yet self-regulation is exactly what President Trump endorsed yesterday after meeting with tech leaders. "There's a belief that there should be tremendous self-regulation," he said, pointing to existing oversight at the Justice Department and the FBI. He offered no coordinated government framework, which is precisely what this moment demands. He has also called fears of runaway AI a "hoax."

The AI lunch attendees all signed an AI agreement that Trump said is  "morally binding." If we are relying on morality to save the day, we are screwed.

The doomsday scenario

Consider what a serious AI failure could look like: drained bank accounts, disrupted hospitals, exposed medical data, scrambled supply chains, blackouts, grounded airlines and trains. There is no light switch and no single plug to pull. Stopping it would take coordination across agencies, deep technical expertise, and leaders who believe the threat is real enough to act before the damage is done.

That capacity lives in agencies that have lost enormous talent. FBI Director Kash Patel has thinned the agency's cyber and counterintelligence ranks through firings and reassignments. The Pentagon, under a defense secretary more worried about testosterone levels and height-weight measurements, has pushed out or lost at least 25 top leaders, with more to come, including the heads of both Cyber Command and the NSA. 

The Justice Department has seen more than 10,000 departures, many in cybersecurity. FEMA, which we'd need if grids fail or water systems are infiltrated, has endured deep staffing and funding cuts.

And what if AI disrupts our health care system and we all get sick? We have a worm-brained health secretary who has overseen a massive loss of brainpower while focusing on anti-vaccine measures instead of outbreaks. Between March 2025 and mid-2026, the CDC's active civil service workforce fell 30 percent, from roughly 12,700 employees to fewer than 8,900.

Those who remain are not necessarily incompetent, but hard-won expertise in how systems break — and how to contain a digital fire — is walking out the door. If AI is as capable as its creators insist, it can presumably see that too. A rogue system wouldn't need to defeat our defenses. It would just need to outpace them.

So which should we fear more? An AI that goes wrong, or an industry with every incentive to keep racing, paired with a government too depleted to intervene? There is no right answer, and that should keep us all up at night.

This article represents the opinions of its writer.

John Casey is a columnist at Raw Story, the Advocate, Daily Beast, and Outsports. He's also the author of Love: The Heroic Stories of Marriage Equality.