A man convicted of manslaughter in Arizona was sentenced to ten and a half years, but an appeals court ruled that an AI-generated video of the victim forgiving him carried undue emotional weight, so the sentence must be reconsidered.
In 2021, Gabriel Horcasitas shot and killed Christopher Pelkey during a road rage incident in Arizona. A jury convicted him of manslaughter.
At the sentencing hearing, Pelkey's sister Stacey Wales played an AI-generated video. In it, a digital likeness of her brother, built with his face and voice, says he forgives his killer and that in another life they could have been friends. Every word was written by her.
The judge loved it and gave the maximum
The judge said she loved the video and handed down ten and a half years, the heaviest sentence available for the charge. Wales told 404 Media that swaying the judge was the whole point, that she wanted to bring the bond between siblings into the room.
The appeals court saw it differently. It ruled that the video, used as a victim impact statement, carried undue emotional weight. The conviction stands; the length of the prison term must be reconsidered. The reasoning, put plainly: a man synthesised from the grave to speak is not evidence, it is effect.
Technology can now make a familiar face say something it never said, and courts have no settled way to weigh what that does to a room.
Why it matters
Sentencing already allows emotion in the room. The question is whether it comes from a living person's breaking voice or from a scripted digital face. Once the victim's own voice can be copied at will, every hearing will have to decide first what counts as real harm and what counts as performance.
Curated from high-quality sources, with concise summaries and key takeaways.
