Web
Analytics
Swifties and George Carlin's Estate Are Gunning for Deepfake AI
Pajiba Logo
Old School. Biblically Independent.

Swifties and George Carlin's Estate Are Gunning for Deepfake AI

By Nate Parker | Celebrity | January 26, 2024

SwiftAI.jpg
Image sources (in order of posting): Getty, Axelle/Bauer-Griffin

On Wednesday, Twitter was inundated with AI-generated deepfake pornography of Taylor Swift at a Kansas City game. One post was viewed more than 45 million times before its removal. The material was shared tens of thousands of times before what remains of Twitter’s security team responded hours later, which is still much faster than they respond to non-billionaires with similar complaints. Online Swifties marshaled a more effective defense, flooding the platform with legitimate photos and videos of Swift’s performances to bury the deepfake porn too deep in search results to be found. Despite their best efforts and Twitter’s eventual removal of accounts posting the material, it’s still available on the platform and other websites. This came just weeks after an audio deepfake of Swift convinced some fans to spend big money on Le Creuset bakeware. The popstar has yet to make a public statement about the harassment, but even someone for whom privacy is little more than a hypothetical has a right to feel violated by this kind of behavior.

Meanwhile, George Carlin’s estate is suing the Dudesy YouTube channel, run by Chad Kultgen and comedian Will Sasso. The pair released an hour-long Carlin comedy special, I’m Glad I’m Dead, two weeks ago. Sasso and Kultgen claim it was written by their AI after all Carlin’s material was fed into its database. Carlin’s daughter Kelly released a strong statement at the time and has now filed suit, saying in part:

“Defendants’ AI-generated ‘George Carlin Special’ is not a creative work. It is a piece of computer-generated clickbait which detracts from the value of Carlin’s comedic works and harms his reputation. It is a casual theft of a great American artist’s work”

The lawsuit’s accusations of copyright infringement are up against claims by AI creators that their theft of intellectual property to “train” their programs is covered by fair use. The concept allows copyrighted material to be used for research, education, and reporting, but is intended for human rather than artificial purposes.

At the same time, Democratic legislators are asking the Department of Justice to investigate reports of political robocalls in New Hampshire designed to stop liberal voters from showing up to Tuesday’s primary. The calls use an AI-generated copy of President Biden’s voice to suggest that voting in the primary would prohibit Democrats from voting in the November election. Robocalls are nothing new, but the ability to reliably mimic a candidate’s voice with open-source software is a new and very sticky complication.

These three cases of fraud, harassment, and copyright infringement are only unusual because they happened so close together, and involved well-known targets financially or legally able to push back. The ease with which AI can be manipulated is a growing issue and one that affects people without these resources every day. Platformer has a solid piece on both the dangers of AI and the frequent warnings legislators have so far ignored. Deepfake porn is used to harass abuse victims, former sexual partners, and whistleblowers. Deepfakes are an obvious threat to fair elections when faked recordings can be used to put candidates in compromising situations or statements literally put in their mouths. Most social media platforms already have policies that prohibit sharing such materials. Those same platforms have also made deep cuts to the security teams responsible for removing it, and aren’t known for their quick responses to anyone worth less than 7 digits. To be fair, bipartisan efforts to move a related bill through Congress are underway, but receiving little support. T. Swift’s online presence and pocketbook might mean legislators finally pay attention, though given recent Fox News conspiracy theories around Swift it’s more likely they’ll make it another joke or attack. Maybe their tune will change when Nancy Mace or Lauren Boebert are inevitably targeted for similar treatment — which would be equally bad, obviously. Even terrible people deserve safety from sexual harassment and revenge porn.