Tackle online abuse with federal AI legislation

OPINION:

For more than two years, I was harassed, threatened and impersonated online by a tech-savvy cyberstalker who, after a long and complex legal battle, is now in federal prison.

That nightmarish experience left me fiercely determined to protect others from online abuse and ensure online abusers are held legally accountable for their actions. Teaming up with fellow survivors and advocates, I worked to advance legislation addressing image-based sexual abuse.

A few months ago, I was thrilled to join a bipartisan group at the White House to watch President Trump sign a milestone bill, the Take It Down Act, into law. Sponsored by Sens. Ted Cruz, Texas Republican, and Amy Klobuchar, Minnesota Democrat, the act criminalizes the nonconsensual sharing of intimate imagery, including artificial-intelligence-generated deepfakes.

The bill’s enactment is a huge win for the victims of online abuse and a critical step toward protecting future generations from abuse’s devastating effects. Yet, given that Congress recently rejected a moratorium on state AI laws, it’s a victory that comes with a caveat: Unless federal lawmakers pass smart, comprehensive federal AI legislation, the Take It Down Act may prove heartbreakingly ineffective.

Here’s why. The Take It Down Act requires online platforms to remove illicit imagery within 48 hours after victims alert them to the images’ presence. Abusers rely on digital communication’s speed and scale to maximize reputational harm to victims, so rapid removal is essential. Still, it can happen only if victims are aware their image is being used. That requires AI-powered tools that can scour the web and quickly detect the images. Unfortunately, a spate of proposed state AI laws may soon create a regulatory patchwork that will make it hard for those tools to work.

In 2023, I left my corporate tech job to launch a company that helps combat malicious deepfakes. As AI became increasingly available to the public, I foresaw a tsunami of new abuse-enabling technologies, and I wanted to do everything I could to protect people from the kind of victimization I had endured. My company’s AI-powered software identifies images and videos that have been manipulated or are fully AI-generated. We also help facilitate the rapid removal of unauthorized content shared across platforms, protecting individual likenesses from digital impersonations.

Still, a tangle of state-level AI laws could bring our work, and the broader fight against online abusers, to a halt. Our tools require AI-powered operations and insights because they are grappling with enormous amounts of complex data and AI-generated deepfake images.

Over the past year, U.S. states have introduced nearly 1,000 different AI-regulation bills, each with different rules and thresholds and distinct reporting, compliance and safety requirements. Many offer different definitions of artificial intelligence. In some states, such as Colorado, my company’s tools will likely be significantly blunted. In many other states, my business could be tied up with various costly, time-consuming legal and oversight requirements.

Taken as a whole, we’re looking at a chaotic regulatory mess requiring teams of lawyers and compliance specialists, not to mention millions of dollars, we don’t have. Meanwhile, criminals and abusers, unconcerned by the laws, will continue harming their primary victims: women and children.

I’ve fought long and hard against online abusers, and I have zero intention of giving up, but if lawmakers are serious about stopping online abuse, they also need to get serious about the difficult task of regulating AI. That starts with recognizing that fragmented, conflicting state laws raise major barriers to innovative AI-powered solutions and, in the case of my work, thwarting and prosecuting online abusers. The U.S. urgently needs a single, national AI-regulatory framework that focuses on stopping and penalizing misuses of AI rather than broadly regulating AI development and innovation.

I understand why the recently proposed moratorium on state-level AI regulation failed; AI clearly needs lawmakers’ attention. Here in Washington state, the Legislature introduced several AI regulation proposals this year. Although I appreciate the efforts, I worry that those bills could have unintentionally hurt our state’s future as a technology leader. More broadly, I worry that a patchwork of often heavy-handed state laws will stifle innovative, law-abiding companies like mine.

I urge legislators at the state and federal levels to meet with and listen to AI startup founders and do the difficult work of crafting thoughtful federal legislation that minimizes AI’s potential risks, maximizes its potential benefits and ensures the good guys have the tools they need to fight the bad guys.

• Melissa Hutchins is the founder and CEO of Seattle-based digital image protection service Certifi AI.