You might use AI to draft an email or research a project, but you are not alone if you feel a little uneasy while doing it. A recent international survey found that a majority of daily AI users in the U.S. feel worried about the technology. It seems that the more we integrate these tools into our lives, the more we question what happens when they get smarter, more autonomous, or more powerful than we intended.
WHAT'S HAPPENING
There is a massive, global tug-of-war happening right now over how to keep AI in check. In California, new laws will force data centers—the physical buildings that house the massive computers that run AI—to be transparent about how much electricity and water they consume. On a federal level, new legislation has been proposed to strictly limit the development of artificial superintelligence, which is an AI capable of performing beyond human ability in almost every way. Meanwhile, global leaders are discussing the need for a formal system—similar to a Cold War hotline—that would allow countries to alert each other about national security incidents involving AI.
Who gets to set the rules?
HOW IT WORKS
To understand why this is suddenly so urgent, you have to look at how these systems are built. Modern AI relies on training models—essentially teaching software to spot patterns in vast amounts of data. This process requires enormous physical infrastructure, known as data centers. These facilities need constant, massive amounts of electricity to run their processors and huge volumes of water to keep those processors cool. As these models get smarter, they need more power, which puts strain on local electrical grids and water supplies.
The other piece of the puzzle is alignment, which is the technical term for making sure an AI actually does what its creators intended, rather than going off the rails. Some researchers are worried about recursive self-improvement, which is when an AI uses its own coding skills to write better versions of itself. If an AI can improve itself faster than humans can monitor it, we lose the ability to steer it, creating a scenario where it might act in ways that are counter to human safety or interests.
WHY IT MATTERS
This is not just about abstract safety; it is about who holds the power. When a handful of private companies build the most capable AI systems in the world, they essentially decide how that power is used. Critics and religious leaders are now warning that we risk letting these labs act like a cartel, where they define the risks and benefits themselves while the rest of us are left out of the conversation.
We are currently in a period of intense adjustment. The technology is clearly useful for everyday tasks, but we haven't yet figured out how to balance that usefulness with the need to keep our resources stable and our societies secure. The big question moving forward isn't just what AI can do, but how we ensure that as it gets more capable, it remains a tool for people rather than a system that makes decisions on our behalf.
Liked this one? The next lands at breakfast.
Every story in tomorrow's AI news, rebuilt in plain English — five minutes, sources linked, free forever.
By joining you agree to receive Article's daily newsletter — unsubscribe in one click. Privacy