At a murder trial that has drawn national attention, prosecutors played two versions of an AI-generated love song that the defendant, Caleb Flynn, created for his mistress. Flynn, a former American Idol contestant, is accused of killing his wife. The songs were discovered on his iPhone by forensic investigator Joseph Wilhelm, who used the GrayKey tool to perform a full file system extraction after other methods failed.
Wilhelm testified that the two versions were in different keys—one upbeat, one somber—and that Flynn had used the software ElevenLabs to have an AI voice sing lyrics he wrote. Earlier in the day, Flynn's mistress read text messages in which he described finishing the song and expressed frustration that the AI had made the vocals "a little extreme." When the recording was played in court, the room reportedly fell into a strained silence as those present tried not to laugh during a murder trial. The judge then adjourned for the day.
The case also highlights how much personal data investigators can pull from a phone. Beyond the songs and messages, Wilhelm obtained location history, sleep data from Flynn's Oura Ring, heart rate data from an Apple Watch, and evidence that Flynn had deleted an app called "AI music song generator" days before the murder. The combination of forensic tools and connected-device data turned an ordinary smartphone into a detailed record of the defendant's activities and state of mind.