Every week brings a fresh cautionary tale about artificial intelligence's darker applications. Convincing fake videos, synthetic voices, manipulated images, deepfakes designed to deceive and defraud. The narrative that accompanies these stories has calcified into something almost religious: technology is advancing faster than our ability to regulate it, the cat is out of the bag, society will simply have to adapt.
This framing is being sold as inevitable. It deserves more skepticism than it is getting.
The latest iteration involves AI-generated romantic profiles targeting vulnerable people on dating platforms. The scam is straightforward and cruel: someone creates a convincing fake persona, builds emotional connection, then requests money. The technology makes it easier. The victims are devastated. Platforms issue statements. Experts shake their heads about the pace of innovation. Everyone moves on, resigned to a future where distinguishing AI fabrications from reality becomes an individual burden.
But acceptance of inevitability is a choice, not a law of nature.
Consider how we've approached other technologies that posed novel risks. When cars became popular, society didn't shrug and accept mass casualties as the price of innovation. We created licensing systems, traffic laws, safety standards. The process was messy and incomplete, but the premise wasn't that cars were too fast and we'd simply have to learn to live with them. We built infrastructure to manage them.
The AI deepfake challenge is genuinely difficult. No question. But difficulty is not the same as inevitability. When tech executives and some researchers insist that regulation will always lag behind capability, they're making a claim about political will, not physics. They're suggesting that democratic systems are inherently too slow to govern emerging technologies. That might be true. But if true, that's a problem with our political systems, not a reason to surrender.
The platform model compounds this fatalism. Social media companies have spent years perfecting a particular rhetorical move: present a problem as belonging to society at large, not to themselves. AI deepfakes target dating apps? Everyone must become more skeptical consumers of digital media. Synthetic voices impersonate loved ones? Families need better verification protocols. The responsibility migrates from those profiting from the platforms to the people being harmed by them.
This is elegant if you're running a platform. It's considerably less elegant if you're someone who has just lost money to a fake persona, or whose image has been weaponized without consent.
There are reasonable debates about what kind of regulation makes sense. Should platforms be required to detect and remove AI-generated content at scale? Should there be authentication systems for videos and audio? Should creators of deepfakes face legal consequences? These are genuinely hard questions involving free speech, technical feasibility, and international coordination. Reasonable people disagree.
But the disagreement should happen in a space where resistance to regulation is treated as a policy choice, not as acceptance of mathematical reality. When tech leaders insist the genie is out of the bottle, they're not describing a fact about bottles and genies. They're expressing a preference for minimal oversight.
Society has choices. We can require platforms to invest heavily in detection tools. We can create liability frameworks that shift incentives. We can invest public resources into competing technologies. We can educate people about manipulation. We can do all of these things imperfectly and incompletely, which is how we typically govern difficult problems.
The deepfake crisis is real. The response should not be resignation. It should be the boring, difficult work of democratic governance applied to a new challenge. That's not inevitable either, which is exactly why we should stop pretending it is.