As I See It: With Deepfakes, Nothing Is Real Except The Consequences
April 15, 2024 Victor Rozek
With an election looming that promises to be conducted with snake-oil ethics, the Congress recently turned its attention from its usual dysfunctional finger pointing to hear testimony about the growing plague of deepfakes.
The usual suspects were represented, in this case Google and Meta, by a spokesman from the industry-funded lobbying group NetChoice. Not surprisingly, they wrapped themselves in the First Amendment and accepted little or no responsibility for what their users post online. They argued that existing laws governing fraud and harassment were sufficient to protect targets (aka victims) of deepfakes.
Standing against the “it’s-not-our-fault” brigade was the mother of a teen who was victimized by the distribution of AI generated nude images. There was, in fact, little disagreement that women and children are the most common victims of deepfakes. A startling statistic was read into the record by Representative Gerry Connolly of Virginia: He referenced a report that claimed 98 percent of all online deepfake videos were pornographic, and that women, girls, (or adults made to look like girls), were depicted in 99 percent of them.
The problem with relying on existing laws to combat AI abuses, explains Wendy Maltz, psychotherapist and sexuality expert (and full disclosure, a family friend), is that “with AI, as with written pornography, there is no crime scene, no sentient victim.” In fact, the entire experience is insensate. The user has to fantasize that it’s real. Essentially, said Maltz, AI porn “digitally dismembers and reconstitutes human flesh for sexual stimulation. It’s no more real than relating to a shiny doorknob.” In other words, it can be legally tricky to identify an actual victim and charge a perpetrator.
People whose lives have been ruined by deepfakes would disagree about the difficulty identifying victims. There have been children who reportedly committed suicide after being targeted by their peers. But there is little to be gained from suing some creep living in his parent’s basement for damages. Hence the focus on AI companies that enable him.
But, since AI regulation is essentially non-existent, deepfakes are proliferating in everything from comedy and satire, to pornography and politics. And your company, where you make your living, could be next.
The first notable attempts at voter suppression were robocalls with an AI generated voice that sounded like President Biden telling voters to not bother voting in an upcoming primary. While most voters would be skeptical of politicians telling them not to vote, the larger concern is the deliberate spread of misinformation and the denial of reality itself. Hany Farid, a professor at the University of California at Berkeley who studies digital propaganda and misinformation says AI creates a “liar’s dividend” in that it provides plausible deniability. When, for example, someone secretly records a politician raging against minorities, or police officers beating a suspect, or provides video evidence of a judge taking a bribe, the immediate defense is that it was AI generated. Thus, the very concept of truth is destabilized and the ability to make critical decisions from a common factual shared reality is undermined.
If you can’t trust what you see, you have to trust the source presenting the information, which further contributes to the tribalism and factionalism already plaguing the country. AI becomes a perfect instrument in a post-truth world: On one hand, AI can create convincing video of people saying and doing the most abhorrent things; on the other, those who are recorded actually saying and doing those things, will deny it and blame it on AI.
However, I believe the problems with AI will not be – and cannot be – solved by legislation. The sources of deepfakes are global and are already being used to influence elections at home and abroad. Kari Lake’s image was used in a deepfake by a group calling itself Arizona Agenda and it was good enough to fool a seasoned reporter. Recently, London’s Mayor Sadiq Khan said that fake audio of him making inflammatory comments before last year’s Armistice Day almost caused “serious disorder.” Even people as powerful and wealthy as Taylor Swift are not immune. As has been widely reported, she has been the victim of musical deepfakes and nude photos that proliferated on the Internet.
In fact, the rich and powerful are themselves raising the alarm. At the recent aren’t-we-special conference of CEOs, and their pet political lackeys, in Davos, Switzerland, AI took center stage. The Washington Post reports that in her opening remarks “Swiss President Viola Amherd called AI-generated propaganda and lies ‘a real threat’ to world stability, “especially today when the rapid development of artificial intelligence contributes to the increasing credibility of such fake news.”
But those who traffic in instability are arguably winning. Greed and malice are powerful motivators and AI is simply a ready tool in their service. The larger problem has always been humanity’s extraordinary gullibility. To a greater or lesser degree our’s has always been a post-truth world because people will believe absolutely anything. From the prospect of hanging with Odin in Valhalla, to burning women for being witches, to Bigfoot and Nessie, or the QAnon claim that Democrats eat babies; facts are superfluous, proof is unnecessary. Only belief counts.
Recall the 39 members of the Heaven’s Gate Cult who committed mass suicide because they were told there was a spaceship hidden in the tail of the comet Hale-Bopp that was coming to scoop them up and return them to the planet of their origin which they called Next Level. There must not have been a Walmart on Next Level because they each packed a suitcase for the journey before drinking the final Kool-Aid.
And so it goes. Discernment cannot be legislated. It is the antithesis of the instant gratification culture. It takes time, work, research, and a heavy dose of skepticism. In a pivotal election year we will be deluged by deepfakes from unknown sources with unknown intentions. In a very real sense, truth, transparency, and consent will also be on the ballot. These virtues are not exclusive to party or ideology, but they will require ethics on the part of information presenters, and a great deal of discernment on the part of viewers.
RELATED STORIES
As I See It: The Other Eight Hours
As I See It: Elusive Connections
As I See It: Entitlement Master Class
As I See It: Greetings, Comrade IT Professional
Let’s be abundantly clear here… Google, Meta, and Twitter (at the time) certainly did not “wrap themselves in the First Amendment” when they extinguished the “Hunter Biden Laptop” story (or anything else that damaged Biden and/or hurt Trump).
Media manipulation was here LONG before AI. The legacy/mainstream media outlets have been practicing this pre 2016. It was (and still is) a conscious effort from the liberal elites that control 90% of mainstream and online information. They’re just not trying to hide it anymore. I need only say three letters N-P-R.