AI is here, and we are in trouble

[Arturo Bandini]

For me AI is like Internet or money. It is neutral. It just depends of the people using it.

Do I have faith that it will be use wisely and for the greater good.

Of course, no ... And that is depressing.

[GMDino]

Ah...I get it now.

https://www.cnn.com/2024/02/19/us/ai-generated-voices-victims-gun-control-cec/

Quote:Joaquin Oliver’s voice echoes through the hallways of Congress on the sixth anniversary of his death.


It comes from his mother’s cell phone outside the office doors of lawmakers on Capitol Hill. The disembodied voice of the 17-year-old, who was a student at Marjory Stoneman Douglas High School, has a message to deliver.


“Six years ago, I was a senior at Parkland. Many students and teachers were murdered on Valentine’s Day that year by a person using an AR-15 assault rifle,” the voice says. “It’s been six years, and you’ve done nothing. Not a thing to stop all the shootings that have continued to happen since.”

“The thing is, I died that day in Parkland,” the voice continues. “My body was destroyed by a weapon of war. I’m back today because my parents used AI to recreate my voice to call you.”



Manuel and Patricia Oliver lost their son, Joaquin, 17, in the 2018 mass shooting at Marjory Stoneman Douglas High in Parkland, Florida. Their organization, Change the Ref, helped launch the new voice message campaign. 
Cody Jackson/AP


The audio is one of six AI-generated voice messages from young people killed by gun violence, part of a new campaign launched last week by two groups, March For Our Lives and Change the Ref, to urge lawmakers to act on gun control.


“My wife and I have been trying to use our voices for the last six years. Nonstop. We have tried almost every single way to approach gun violence in a way that people will pay attention. We haven’t been very successful,” says Manuel Oliver, who founded the Change the Ref advocacy group in memory of his son after the 2018 shooting in Florida.


“So we decided, you know what? Let’s bring the voices of our loved ones. Let’s bring the voice of Joaquin.”


One man listened to his son’s voice over and over before signing off
The new campaign’s website, called The Shotline, invites people to listen to the voice messages, enter a zip code and send calls to members of Congress.


The campaign launched the same day a mass shooting at a Super Bowl victory parade in Kansas City killed one person and wounded more than 20 others, including children. There have been 50 mass shootings in the United States so far this year, according to the Gun Violence Archive.



Uvalde shooting victim Uziyah "Uzi" Garcia. 
Mitch Renfro


Uzi Garcia, 10, who died in May 2022, is one of the six children featured in the campaign.


“I love video games, telling jokes and making my friends laugh and jumping on the trampoline with my family,” Uzi’s AI-generated voice says in his message.  “I’m a fourth grader at Robb Elementary School in Uvalde, Texas. Or at least I was when a man with an AR-15 came into my school and killed 18 of my classmates, two teachers, and me. That was almost two years ago. Nothing has changed. Even more shootings have happened.”


Uzi’s father, Brett Cross, told CNN he spent hours going through old playground videos on his phone to get the right audio of Uzi’s voice. He says he worked with the campaign’s technology team to make sure the pitch was just right and he listened to Uzi’s AI voice over and over before he signed off on its use.


“It was bittersweet because we get to hear his voice again,” he says. “But he should be here to speak for himself, and he’s not. So we have to get more and more creative to get these politicians to listen to us.”


Cross says he understands that employing the voices of dead children to plead for gun control bothers some people. He says critics on social media have described the campaign as unethical and “ghoulish” and have accused him of using his son as a ”pawn” to push a gun control agenda.

Brett Cross says he's unfazed by critics who are against the use of AI-generated voices of children killed by gun violence. "It was bittersweet because we get to hear his voice again,” he says of his son. 
Nuri Vallbona/Reuters


But he told CNN he’s unfazed by the criticism.


“If you think it’s uncomfortable hearing my son’s voice after he’s passed, imagine what it’s like to be us — to live with this every day,” he says.


AI-generated voices can raise legal and ethical concerns
Since the campaign launched, more than 54,000 voice calls have been sent so far to lawmakers, according to a tally on The Shotline site.


The campaign features four other victims of gun violence:  Ethan Song, 15, of Connecticut, who died in an accidental shooting in 2018; Akilah Dasilva, 23, who was killed in a 2018 mass shooting in Tennessee; Mike Baughan, 30, of Maryland, who died by suicide in 2014; and Jaycee Webster, 20, who was shot and killed at his Maryland home in 2017.


More than two dozen other parents have contributed audio of their children’s voices to the campaign for potential use in future calls, Manuel Oliver says.


The campaign comes as rapid advances in AI technology have made it easier to recreate people’s likenesses and voices, allowing scammers and other bad actors to send fake and manipulative messages.

Uzi Garcia was a victim of a May 2022 massacre that left 19 students and two teachers dead at Robb Elementary School in Uvalde, Texas. 
Jordan Vonderhaar/Getty Images


The Federal Communications Commission allows political prerecorded voice calls made to landlines, even without prior consent. But earlier this month, it  announced that robocall scams using AI-generated voices are a violation of telecommunications law. The announcement came after an AI-generated robocall that sounded like President Biden circulated to New Hampshire voters ahead of the state’s January primary election, telling them to stay home.


The Shotline’s messages are sent to lawmakers’ landlines and make it clear that the voices are AI-generated, Oliver says.


One expert told CNN the parents are likely not breaking FCC law because they’re transparent about their use of AI and their aim is not to defraud.
“If you evaluate this call from that perspective, it’s clear up front it’s AI generated, so if they have permission from the families of the voice they’re using, and they’re not violating the TCPA (Telephone Consumer Protection Act) in some other way, it would be fine,” says Alex Quilici, CEO of YouMail, a robocall-blocking service. “Think about the old ads that said ‘celebrity voice impersonated’ — we were all good with those.”


But for some, using AI-generated voices of dead people in a gun control campaign may raise ethical concerns.


“It does seem to kind of straddle the line between a good use of AI and something that’s questionable,” says Robert Wahl, an associate professor of computer science at Concordia University Wisconsin and an expert on the ethics of AI technology.


Creating AI voices is a nuanced process that involves numerous adjustments to make sure the inflection, timing and pitch are as close as possible to the original voice and don’t sound robotic, he says.


“It’s interesting because Hollywood has been doing this kind of thing for a while with recreating dead actors,” he told CNN. “And of course the technology is evolving almost on a daily basis. But I think it’s nice to be able to recreate a voice, as long as it’s OK with the immediate family.”


Manuel Oliver says he understands some people’s discomfort with the campaign’s AI-generated voices. He says some parents who’ve lost children to gun violence were reluctant to take part in the project.


But his goal is to jolt members of Congress and others into action on gun laws. And he says if it bothers some people — well, that’s the point.
“If bringing the voice of a victim alive using technology in a transparent way … If that makes you feel uncomfortable, but you’re OK with kids getting shot at parades and schools, then something’s wrong with you,” he says.


“No one should tell me what the hell uncomfortable is … because I can tell you what feeling uncomfortable is. It’s not being able to see my son ever again.”

[Sociopathicsteelerfan]

(02-20-2024, 09:45 PM)GMDino Wrote: Ah...I get it now.

https://www.cnn.com/2024/02/19/us/ai-generated-voices-victims-gun-control-cec/

Wow, I guess waiving the bloody shirt became too passé.  Also, congrats on heavily politicizing a topic that should be solidly non-partisan.  Must be all you got.  Sad.

[GMDino]

According to this article more white collar jobs will be in jeopardy than anything else.

We could see a replay of the 80's where all the "middle men" and middle management got outed (for the sake of profits and to avoid job duplication...of course) and those same companies moved so much of their production overseas taking away the better paying blue collar jobs already.

My job, for example, could easily be done with a chatbot...except the bot would have zero idea about what to sell without a human entering all the information on a daily basis.

Frankly I've been looking for an AI solution to some of the more mundane, week-to-week duties I have.  that would free me to do more of the things only a human can do.

[Goalpost]

Google's Gemini feature is majorly screwed up.

[CKwi88]

(02-21-2024, 10:31 AM)GMDino Wrote: According to this article more white collar jobs will be in jeopardy than anything else.

We could see a replay of the 80's where all the "middle men" and middle management got outed (for the sake of profits and to avoid job duplication...of course) and those same companies moved so much of their production overseas taking away the better paying blue collar jobs already.

My job, for example, could easily be done with a chatbot...except the bot would have zero idea about what to sell without a human entering all the information on a daily basis.

Frankly I've been looking for an AI solution to some of the more mundane, week-to-week duties I have.  that would free me to do more of the things only a human can do.

I can definitely see this. As a white-ish collar worker, about 95% of my administrative duties could easily be replaced with a machine. I've been able to automate much of my work just through custom Outlook signatures. Imagine if I had the know-how to write some simple if "x" then "y" code for the computer to do it for me. 

Thankfully, AI can't quite replace my skill-based duties. Although even on that front (language interpretation and translation) machines have made great strides in the past years. I've worked with computer-assisted translation software for the past 7-8 years, and it used to be laughably bad. Bad to the point that even if I ran something through an AI translator it would be just as much work to fix the translation as it would to just do the translation from scratch. Now AI is even able to (with enough prompting at least) even accurately translate judicial concepts from Brazil into state-specific equivalents from here in the US. 

Then again, if we ever make an AI that masters language to that level I'm of the opinion that will mean we've created AGI which will not have me worrying for my job but our species' very existence.

[NATI BENGALS]

(02-27-2024, 10:22 AM)Goalpost Wrote: Google's Gemini feature is majorly screwed up.

Saw an interesting discussion on CNBC this morning about how damaging that could be.

Google has been where people go to get facts. The correct information that is historically accurate. Now you have their Ai spitting out misinformation.

Teachers use the stuff. If it is relaying bad information that can make its way in to text books and class rooms. And/or people just using it on their own and accepting what it says as accurate.

We are already at the acceptable upper limit of dumbasses. I’m not sure how the world would manage with that number growing exponentially

[Belsnickel]

(02-27-2024, 10:40 AM)CKwi88 Wrote: I can definitely see this. As a white-ish collar worker, about 95% of my administrative duties could easily be replaced with a machine. I've been able to automate much of my work just through custom Outlook signatures. Imagine if I had the know-how to write some simple if "x" then "y" code for the computer to do it for me. 

Thankfully, AI can't quite replace my skill-based duties. Although even on that front (language interpretation and translation) machines have made great strides in the past years. I've worked with computer-assisted translation software for the past 7-8 years, and it used to be laughably bad. Bad to the point that even if I ran something through an AI translator it would be just as much work to fix the translation as it would to just do the translation from scratch. Now AI is even able to (with enough prompting at least) even accurately translate judicial concepts from Brazil into state-specific equivalents from here in the US. 

Then again, if we ever make an AI that masters language to that level I'm of the opinion that will mean we've created AGI which will not have me worrying for my job but our species' very existence.

I have written so many macros to do some of the repetitive and mundane tasks I have done in my job as well as created scheduled queries to pull the data. A week of work and it changed a 3-hour a day task into 15 minutes and just a handful of clicks.

[Sociopathicsteelerfan]

(02-27-2024, 09:59 PM)Belsnickel Wrote: I have written so many macros to do some of the repetitive and mundane tasks I have done in my job as well as created scheduled queries to pull the data. A week of work and it changed a 3-hour a day task into 15 minutes and just a handful of clicks.

You are not lying at all.  I was helping my, at the time, direct superior with a DOJ audit.  I'll preface by saying I am inherently lazy, especially outside of work, so I am all about maximizing efficiency.  After helping for a few days I developed a routine that reduced the process tremendously.  It got to the point that I was providing what took some people several work days to complete in less than two hours, using similar methods.  It definitely helped me promote not long after.  Sadly, I tried to teach this to others with less than preferred results.

[*TheLeonardLeap*]

(02-18-2024, 10:40 AM)BigPapaKain Wrote: Well, anyone who grew up liking sci-fi knew humanity was doomed in one of two ways - glad to see we went the route that involves both options.

[Sociopathicsteelerfan]

(02-28-2024, 02:34 PM)TheLeonardLeap Wrote:

Tit's, I mean, it's all over for us.