Donald Trump’s big bill of tax breaks and spending cuts runs about 900 pages long. It’s the perfect vehicle for trying to sneak in provisions that cut out special treatment for certain industries.
Thankfully, a 450-word section that would have given technology companies 10 years to develop their artificial intelligence capabilities free of state government oversight got sniffed out. The House-passed proposal was killed by a vote of 99-1 this past week in the Senate following a firestorm of protest from a bipartisan group of governors and other state officials.
If Congress is wise, it will not resurrect this carve-out. Lawmakers only have to think back to the protections they gave Big Tech when the internet was first emerging as a commercial enterprise almost three decades ago. The decision to exempt online platforms, such as Facebook, X and other social media sites, from the nation’s libel laws has helped them decimate the business models of traditional media, coarsen political discourse and turn the internet into a cesspool of slander, rumor, lies and cruelty.
The argument for letting artificial intelligence companies grow unfettered echoes what was being said when Section 230 of the Communications Decency Act was passed in 1996. Back then, the argument was that the internet was in its infancy, and that its potential would never be realized if the companies that were commercializing its use had to worry about everything posted on their platforms. That exemption, though, gave them a huge competitive advantage over the traditional media – newspapers, radio and television – whose advertising they aimed to take away. This exemption helped turn these infant companies in a matter of a decade or so into huge, immensely rich monopolies.
Proponents in Washington for bestowing legal immunity to artificial intelligence warn that too much U.S. regulation will stifle innovation and help China achieve global dominance in this newest technological frontier.
Although there are many developments within AI that are exciting, there are also plenty of dangers. The states have stepped up, in the absence of any congressional desire to regulate the tech industry, to try to impose commonsense safeguards.
This year alone, according to the National Conference of State Legislatures, state lawmakers around the country have filed more than 1,000 pieces of legislation related to AI. Twenty-eight states and the U.S Virgin Islands have adopted at least 75 of these measures.
The congressional carve-out would have blocked enforcement of any of those laws and the adoption of any more for a decade, leaving AI companies to make money without consequence on deepfakes that impersonate politicians and celebrities and turn children into virtual porn stars. Gone would be the newer protections provided to entertainers, authors and artists for the commercial rights to their creative work. And traditional media, some of which are suing over AI’s theft of their material and its unfair business practices, would face an existential threat that’s possibly even greater than was the birth of the internet.
Those are only the dangers, however, that we know about now. Who knows what harms may emerge that no one has yet imagined?
Ten years is a lifetime in an industry that is progressing this exponentially fast. Giving AI a decade long free pass would entrench it so much, enrich it so greatly, and embed it so deeply in daily life that whatever harms it created would be about impossible to reverse.
Rather than trying to block the states from regulating AI, Congress should put its energy into developing legislation that takes into account the states’ best ideas and applies those safeguards nationwide. If a 50-state patchwork of regulations is too onerous for the AI industry, the answer is not to lift oversight but to make it uniform and meaningful.
