A template for AI policymaking

Artificial Intelligence holds immense promise to benefit our economy and nation. The advantages of this emerging technology are hard to overstate. However, it is also critical to acknowledge the real unintended consequences that cause tangible harms. Finding a way to enable innovation while also reining in bad outcomes is a difficult but necessary task for crafting a policy framework for AI.

There is a path forward on these nuanced questions, but too often legislators rush to introduce bad proposals that will only harm innovation while not mitigating the tangible harm. A recent bill signing by President Donald Trump, however, shows a template for smart AI policymaking.

Over the past few years an epidemic of deepfakes and revenge porn has permeated the internet culture. This practice takes advantage of vulnerable individuals or exploits bad decision mostly made by teens and young adults. In a rare moment of bipartisan unity, Congress took a decisive stand against one of the darkest misuses of technology. The Take It Down Act is a sweeping law targeting non-consensual intimate imagery, including AI-generated deepfakes.

The legislation, championed in the Senate by Commerce Committee Chairman Ted Cruz, R-Texas, and in the House by Rep. Maria Elvira Salazar, R-Fla., criminalizes the publication of non-consensual sexual content, including AI-generated nudes. It also mandates that social media sites and other online hosts remove the offending content within 48 hours of receiving a request to do so from a victim.

Over the past several months support for the proposal snowballed—mainly because Melania Trump used the bully pulpit of the first lady to champion this legislation, making it a top priority of hers since Inauguration Day. It was Melania who invited a victim of deepfake images to sit with her during the president’s joint address to Congress in March. At the speech, President Trump explicitly endorsed the legislation, stating, “I look forward to signing it into law.”

In early March Sen. Cruz and Rep. Salazar joined Melania Trump for a roundtable featuring a number of victims of deepfake and revenge porn. The event was a rare example of the first lady specifically calling on Congress to pass legislation. The event was at times very emotional with several victims (mostly teenagers) going into detail about their harrowing experiences, making clear the real-world stakes of this crisis.

In an unusual moment of timely action, Congress responded quickly. The Senate passed the bill unanimously, and the bipartisan Problem Solvers Caucus added its endorsement, reflecting rare cross-aisle agreement on the need to protect victims from digital exploitation. On April 28, the House of Representatives also passed the bill by a vote of 409-2.

When President Trump signed the bill into law on May 19, Melania joined her husband for the ceremony. “Artificial Intelligence and social media are the digital candy for the next generation—sweet, addictive, and engineered to have an impact on the cognitive development of our children,” the first lady said. “But unlike sugar, these new technologies can be weaponized, shape beliefs, and, sadly, affect emotions and even be deadly.” In a rare move President Trump invited her also to sign the law she had worked so hard to achieve.

This legislative accomplishment demonstrates how Congress can bring about thoughtful proposals to mitigate the real harms of emerging technologies while not trampling on free speech and economic opportunity. We are still in the nascent stages of the growth of exciting technologies. Instead of taking clumsy actions that will have dramatic impacts, this bespoke policy remedy serves as a template for success. Let’s hope that that over the next several years more of our discourse over AI governance will follow this balanced framework.