Avatars of activism

On Sept. 14, worshipers at Prestonwood Baptist Church in Plano, Texas, and Dream City Church in Phoenix, Ariz., received an unusual message—an exhortation from the late Charlie Kirk, who was assassinated on Sept. 10. But the message wasn’t from Charlie himself; it was an AI-generated response using a simulation of Charlie’s voice and cadence.

In the one-minute message, AI-Charlie assured everyone that he was fine; “my soul is secure in Christ,” so “don't waste one second mourning me.” AI-Charlie exhorted the congregation, “Do not let this violence divide us further. The enemy wants chaos, fear, and retaliation. Don't give it to them. Instead, double down on truth. Double down on courage. Double down on your faith and on your families. That is how you honor me.” He concluded by reminding them that “America is worth it. Free speech is worth it. Fighting for the unborn, for families, for sanity in a culture gone mad. It is all worth it. So dry your tears, pick up your cross, and get back in the fight. Do it with joy. Do it with strength. And never ever let evil think it won.”

The clip is another phase in the growing controversy over the use of AI avatars of the deceased, from the use of a combination of AI, CGI, and a body double to put a likeness of Grand Moff Tarkin in the Star Wars film Rogue One (Tarkin was originally portrayed by Peter Cushing who died in 1994) to a family in Arizona who created an AI video of a murdered loved one to extend forgiveness to his killer in court.

More recently, former CNN anchor Jim Acosta conducted an interview with an AI-generated avatar of Joaquin Oliver, a 17-year-old victim of the 2018 Parkland school shooting. The five-minute interview with AI-Joaquin was followed by a ten-minute segment with Joaquin’s father Manuel, who worked with tech companies to create the AI version of his dead son as a part of his gun control advocacy group Change the Ref.

The interview with AI-Joaquin was criticized from both left and right as “creepy,” “ghoulish,” and “a grotesque puppet show.” The most recurring criticism was the use of a dead boy as an avatar for a particular political agenda. Joaquin’s parents, Manuel and Patricia, have publicly stated that they created AI-Joaquin in order to “keep his voice alive” and “fight for a world without gun violence.”

This, of course, is not new, and grieving families across the political spectrum have used the image and memory of their deceased loved ones to advocate for policies they believe would have prevented their deaths (over the last eight years, Joaquin’s image has appeared as a cardboard cut-out at baseball games and on murals in order to raise awareness about gun violence).

But the use of AI takes this practice to another level. And the initial reaction indicates that there may be a backlash against this AI escalation. While most Americans have deep compassion for those who have lost loved ones, and are even willing to tolerate political activism flowing from their loss, the obvious attempt to wield the death of the loved one in such a manipulative manner strikes many as beyond the pale. For example, the transparent attempt by Acosta to “humanize” AI-Joaquin in the second half of his interview seemed particularly manipulative, even if ham-fisted because of the technological limitations. AI-Joaquin talked about his favorite basketball team (the Miami Heat: “They have such a passionate fan base and an exciting playing style”) and Star Wars character (“Luke Skywalker is a classic hero. His journey from a farm boy to a Jedi is so inspiring.”). Most people already dislike guilt trips; how much more guilt trips by AI avatars.

The use of such AI characters as avatars for activism highlights the blurring of the lines between the real and the virtual that our technology represents. Acosta’s conversation with Manuel Oliver demonstrated the tension. On the one hand, Oliver insisted that he is in no way “trying to bring his son back.” “This is AI,” he says. At the same time, he commented that, with respect to the victims of the Parkland shooting, “We’ve heard from the parents. We’ve heard from the politicians. Now we’re hearing from one of the kids.”

But we’re not. No matter how much Acosta might feel like he was “really communicating with him,” the image on the screen was not Joaquin, but AI-Joaquin, a simulacrum built by programmers who combined LLM’s with images, texts, and videos from Joaquin’s life. No matter how much his father might insist that “we’re finally able to add Joaquin’s presence in a more direct way to our movement, to his movement,” Joaquin was and is not present.

The same is true of AI-Charlie Kirk. One pastor may claim that “this was Charlie’s response” to a question from a little girl. But it wasn’t. Another pastor may invite his congregation to “hear what Charlie is saying to us regarding what happened to him,” but we’re not. Charlie and Joaquin are both dead, their lives tragically ended by evil men, and no amount of technological necromancy can change that.

The Bible tells us that human beings are made in the image of God. The question is whether we are truly honoring that image when we construct an image of an image and get it to speak? Or in our desire to extend our lost loved ones, are we losing their humanity? Are we losing our own?