Police arrest high school athletic director for deepfaking principal's voice

Cal Jeffrey

Posts: 4,190   +1,430
Staff member
In brief: Baltimore police arrested a high school athletic director after he allegedly created an AI-generated recording of the school's principal making racist and antisemitic remarks and circulating it on social media. He faces charges, including theft, stalking, disruption of school operations, and retaliation against a witness.

According to NBC News, police arrested former Pikesville High School athletic director Dazhon Darien as he was attempting to board a flight to Houston, Texas. Transportation Security Administration officials detained Darien over the packaging of a declared firearm. When local police arrived, they arrested him after running a background check and finding an outstanding arrest warrant.

Darien allegedly posted the now-deleted audio clip to multiple unnamed social media sites in January. The primary voice in the audio clip sounded like Pikesville High's principal, Eric Eiswert. It appeared to be a conversation with an unnamed individual.

The police report notes that in the course of the dialog, Eiswert says that African-American students couldn't "test their way out of a paper bag." He also made slanderous comments about people in the Jewish community and two school employees who "should never have been hired."

Eiswert was temporarily relieved of his duties after the audio went public. He received high volumes of venomous messages online, and the school received countless phone calls from outraged community members.

Eiswert told police that he believed Darien fabricated the audio using AI in retaliation for his contract not being renewed due to poor work performance, alleged mishandling of school funds, and for terminating a long-time coach without approval in 2023. He also said he believed Darien was tech-savvy and knowledgeable about using AI tools.

"[Darien was] technologically savvy and believed he was familiar with AI," Eiswert told police.

Detectives found evidence that Darien had accessed the school's network several times in December and January to search for "OpenAI tools." Investigators subsequently linked him to multiple paid OpenAI accounts.

The FBI contracted a university professor and forensics expert to analyze the recording. He found that the file contained traces of AI-generated content. The architect of the recording also added a background track of ambient noises to make it sound more natural. A second analysis by the University of California, Berkley, concluded that someone had manipulated the audio file.

Image credit: Shuttersv

Permalink to story:

 
So I'm just gonna throw this out there as I've done a considerable amount of work with AI. If this guy was going to make an AI recording with this guy's voice then he had to have A LOT of his real voice already recorded. You don't just "make an AI voice recording cause you're tech savy." I find it more likely that this guy is actually a racist and using this to get out of being caught as a racist than some guy using AI to mimic his voice.
 
ANY kind of impersonation of any kind needs to be treated as harshly as possible to prevent it from becoming completely out of control. As it is any kind of recording, etc. that would normally be used as evidence is now under specious and it will become the obligation of the prostitution to prove it's origin and authenticity, which in most cases will be nearly impossible to do.
 
So I'm just gonna throw this out there as I've done a considerable amount of work with AI. If this guy was going to make an AI recording with this guy's voice then he had to have A LOT of his real voice already recorded. You don't just "make an AI voice recording cause you're tech savy." I find it more likely that this guy is actually a racist and using this to get out of being caught as a racist than some guy using AI to mimic his voice.
Except for the fact that the perp was recorded looking up openAI at work, was linked to multiple paid openAI accounts, and the FBI found evidence the audio was AI generated.

Perhaps you should check to see if you yourself hold some racist tendencies, to the point you are willing to throw out evidence to support your narrative? You have no basis for saying the victim was actually racist other then your feelings, and your "work with AI" which is anecdotal at best.

You could also review this article made by this site called Tecshpot. Microsoft AI, built on open AI, can make a convincing deep fake on 3 seconds of audio. Your pre conceived notions on AI are incorrect and you should do some more research before making slanderous claims.

https://www.techspot.com/community/...ice-after-analyzing-a-3-second-sample.278638/
 
Except for the fact that the perp was recorded looking up openAI at work, was linked to multiple paid openAI accounts, and the FBI found evidence the audio was AI generated.

Perhaps you should check to see if you yourself hold some racist tendencies, to the point you are willing to throw out evidence to support your narrative? You have no basis for saying the victim was actually racist other then your feelings, and your "work with AI" which is anecdotal at best.

You could also review this article made by this site called Tecshpot. Microsoft AI, built on open AI, can make a convincing deep fake on 3 seconds of audio. Your pre conceived notions on AI are incorrect and you should do some more research before making slanderous claims.

https://www.techspot.com/community/...ice-after-analyzing-a-3-second-sample.278638/
I, too, have used openAI extensively and have even paid for it. I've also been wronged by people who have been caught on tape saying some pretty stupid stuff. I'm not saying what happened is impossible, just that it seems more likely that someone said something stupid and doesn't want to lose their career because they got caught.
 
Back