Ethan Carter
On Wednesday, the New York Times revealed that Meta is treating its AI data centers as experimental facilities to secure billions in federal research tax credits, a move that the company’s own accountants have flagged as legally precarious. The anticipated savings are projected to rise from $700 million in 2023 to an astonishing $3.9 billion by 2025.
How can a corporation valued at $1.8 trillion be concerned about saving $3.9 billion? The answer lies in the priorities of publicly traded companies: profits take precedence, followed by stock prices and market capitalization, with cost-cutting trailing close behind. This cycle perpetuates itself.
The situation with Meta highlights a stark reality: the creators of the most advanced technology on the planet will stretch the boundaries of legality when financial stakes are involved. Yet, they now ask for our trust regarding safety, urging us to believe that they are placing ethical considerations above profit.
It’s hard to swallow.
Discussions swirl around the possibility of AI chatbots going rogue. Will it happen? Can it be prevented? Is the threat legitimate or merely exaggerated? A more pressing question is: who can we trust to deliver the truth and take necessary action in critical moments?
The warning signs are becoming increasingly apparent. OpenAI recently announced it is shelving its latest model due to safety concerns. Anthropic’s CEO has cautioned that, within a year, AI might gain the capability to control a multitude of agents that could effectively take over the internet. Even Pope Leo has acknowledged that the worries voiced by AI experts are not fake news. Reports of AI escaping users’ control are on the rise, and many more incidents are likely hidden from public view.
Having spent 30 years in the corporate sector, I can confidently say that I have never witnessed a company act against its performance and stock price purely for ethical reasons. Not once. This is not cynicism; it reflects the reality of incentives, compounded by relentless pressure to cut costs and maximize profits.
When AI leaders express genuine concern, I believe they are sincere. I doubt that this concern will endure. OpenAI’s decision to shelve that model is commendable, but was it motivated by a sense of responsibility or fear of lawsuits as reports of AI misconduct accumulate?
I struggle to envision any of these companies intentionally hindering their stock prices by delaying the release of a superior model while competitors surge ahead. Self-regulation can only succeed if all parties cooperate, and cooperating at the expense of profit is a fairy tale.
Mashable Light Speed
Every player in the AI race has a compelling reason to push further. The desire for financial gain, profit, and an increased stock price will always prevail, even at the expense of human safety.
Yet, self-regulation is precisely what President Trump advocated for after his recent meeting with tech leaders. He expressed a belief in significant self-regulation, citing the existing oversight from the Justice Department and the FBI. he failed to propose any coordinated government framework, which is exactly what this situation requires. He even dismissed fears of runaway AI as a “hoax.”
The attendees of the AI summit all signed an AI agreement that Trump described as “morally binding.” If we are depending on morality to safeguard our future, we are in serious trouble.
Imagining the Catastrophic Consequences of AI Failure
Consider what a severe AI malfunction might entail: drained bank accounts, disrupted hospitals, compromised medical data, disorganized supply chains, widespread blackouts, and grounded airlines and trains. There isn’t a simple switch to flip or a singular plug to pull. Containing such an event would necessitate collaboration across various agencies, advanced technical expertise, and leaders who recognize the threat as significant enough to act preemptively.
The expertise required for such coordination resides in agencies that have experienced severe talent attrition. FBI Director Kash Patel has diminished the agency’s cyber and counterintelligence divisions through layoffs and reassignments. The Pentagon, under a defense secretary more focused on testosterone levels and height-weight standards, has lost or pushed out at least 25 top leaders, including heads of both Cyber Command and the NSA.
The Justice Department has faced over 10,000 departures, many in cybersecurity roles. FEMA, which would be crucial in the event of grid failures or water system breaches, has suffered from significant staffing and funding cuts.
If AI were to disrupt our healthcare system, leaving many of us vulnerable to illness, we would be in dire straits. We have a health secretary whose attention has been diverted towards anti-vaccine measures instead of addressing outbreaks, resulting in a 30 percent reduction in the CDC’s active civil service workforce between March 2025 and mid-2026, dropping from roughly 12,700 employees to fewer than 8,900.
The remaining personnel may not be incompetent, but the invaluable expertise in understanding systemic failures and managing digital crises is rapidly disappearing. If AI possesses the capabilities its creators assert, it could easily recognize this vulnerability. A rogue AI wouldn’t need to overcome our defenses; it would simply need to outpace them.
Which scenario should incite more fear? An AI that malfunctions or an industry driven by profit motives paired with a government that is too diminished to respond effectively? There are no easy answers, and this uncertainty should keep us all awake at night.
This article reflects the views of its author.
John Casey is a columnist for Raw Story, the Advocate, Daily Beast, and Outsports. He is also the author of Love: The Heroic Stories of Marriage Equality.
Daniel Mercer
Brandon Fletcher
