Something like that could also be a tool for the socially awkward to practice with though. Depends how people use it.
@BoomerSpooner - It might lead to false expectations. They'll never be rejected by an A.I character.
@Mike403 I don't believe that, the not needing real human interaction bit. Do you stop interacting with your friends when you make new ones? Do you only react with people who are carbon copies of each other, no variety at all?
AI is an amazing tool and the potential for growth is ever increasing. But AI cannot give in person hugs. It cannot feel human emotions and empathize, it can only simulate them. I imagine interacting with only AIs would be similar to interacting with only sociopaths - people who do not feel their own emotions but wear masks 24/7 to act like they believe people want them to. As someone who has long masked to hide my own neurodiversity, I know how much of a negative impact that masking can have on genuine emotional interactions.
At some point in the far future AIs and human biology will likely need to merge to some degree in order for AI to progress to that level of human interaction. When that happens, we'll be looking at a new evolution of the species, not merely a tool to be used.
It might lead to false expectations.
Humans never needed technology for that. The concept of marriage has been providing that for centuries.
@cuddlefaery - There's also the potential for misuse. Scammers can use the technology on dating websites so they can cast a wider net especially since it can create unique novel images that can't be reverse image searched.
@Mike403 any tool can be misused. ANY tool. A pencil can stab someone, ruin their reputation through their own words or that of others, splinter and blind someone, sign a death warrant. It can also draft great works of literature, be used by school children learning to write, sign a birthday card of a loved one, and so on.
Do we stop using all tools because of their potential for misuse? Or do we educate on safe use, put in safeguards (legal, technological, and otherwise) so that the tool can continue to be utilized for good while minimizing the harm?
A personal computer can open doors of opportunity to new worlds, provide education and employment, create accessibility for people who face barriers, connect lonely people with communities where they feel included.
It can also be used to scam, buy and sell illegal goods, bully others on social media, provide safe havens for harmful content like child pornography, or even just leave people vulnerable to being targeted.
Do we malign all personal computers because of the potential for harm? Or do we educate people on how to use the tool safely? Limit access to those who would use them to cause harm? Implement consequences for misuse?
AI romantic partner apps / websites are not new. I created a thread about one of them (Replika) in October 2022. By then they had already been around for a couple of years but I had just heard of them on a popular podcast.
https://www.cuddlecomfort.com/forum/discussion/15187/thoughts-on-replika#latest
I also created a follow up thread in February 2023 about how the company behind Replika had pushed an update out to all users that dramatically impacted the AI’s personality and that update was leading to suicidal ideation among many vulnerable users of the app.
https://www.cuddlecomfort.com/forum/discussion/16188/replika-recklessly-exploiting-lonely-people#latest
Abigail Algo did two excellent YouTube videos investigating and exposing the corruption and exploitation in the company behind Replika.
Replika was founded by a Russian immigrant with ties to not just one but two Russian Oligarchs who are on the U.S. State Department “naughty list” (officially known as the sanctions list). The company behind Replika, Luka, is named after the son of one of the Oligarchs.
If the AI doesn't take an hour to get ready when we were supposed to be somewhere 15 minutes ago then I'm signing up.
@JohnR1972 - As long as there's a demand for it, another company will pop up to give people what they want like what just happened. The internet has been full of porn almost as long as the internet existed.
GirlfriendGPT is a pay service costing $15/month for 5,000 messages per month or $35/month for 20,000 messages per month. Images count as 10 messages. NSFW/Erotic Role-Play included.
@Mike403 These type of apps will continue to proliferate and advance in sophistication. In some cases, I can see them being truly beneficial to people as they can “redirect” negative thoughts to more positive topics and/or help reframe the way people thing about things / people / experiences.
However, many of the companies behind these apps are NOT benevolent. Like Facebook, they will offer “services” at no to low cost then collect and sell your data. Just assume anything you share with an “AI friend” is being sent to someone just like Mark Zuckerberg who will find a way to monetize that information.
Emotionally vulnerable people need to be especially careful with these apps as your “friend” may start suggesting products, movies, books, causes, and/or political candidates you might want to check out.
Having the undivided attention and unconditional love of a “virtual friend” is a powerful connection and if the virtual friend starts trying to influence someone, they will know that person so well they will be able to do it in subtle yet persuasive ways.
@JohnR1972 - I may be lonely and too socially awkward to get a girlfriend but I will never resort to a fake relationship with an A.I. girlfriend.
I've used ChatGPT for things that can be beneficial such as coding help for a game that I was working on that I put on the back-burner for now, but it was fun creating something at the time. Here's my thread about it:
https://www.cuddlecomfort.com/forum/discussion/16851/game-development
@TxTom there is a reason people in capitalist countries are doing better. Economic hit man details why in this podcast: https://open.spotify.com/episode/7CADxUmtqm3vtjlqGZKnUO?si=YSVLXoU4SZm6JnsXi5S2mA
@Parad0x make a great point! Dataannotation though is changing the foundation of AI’s ability to be creative
OpenAI is now working on on a text prompt to video A.I.
https://openai.com/sora
Thanks, @coregulate I'll give that podcast a listen this weekend.
A while ago, I asked chat gpt about the movie Her. I asked can you be as good as Samantha. The responses where brilliant!
@Mike403 that will be a great advancement for creative people who use digital visuals, as well as enhancing accessibility for people who struggle with text.
@cuddlefaery - Somebody can use it for malicious purposes such as accuse you of something and create fabricated "evidence". You can't trust anything any more. Not even videos. Deepfake A.I. has already ruined lives.
What's the solution, @Mike403 ?
Gone are the days of needing a skilled photoshop user to manipulate images. That's what makes A.I. image generation so scary. Anyone can do it in a matter of minutes, so there is more opportunity for malicious content to spread if it's not kept in check.
https://www.technologyreview.com/2021/02/12/1018222/deepfake-revenge-porn-coming-ban/
Now the technology is moving onto creating videos.
Creative destruction, that's all it is. It'll be fine.
You know "they" said the same thing about computers killing writing skills? And yet people still write, just in more formats than before for different contexts. Or that digital photography would kill the art and soul of photography - and instead it has encouraged more people to learn photography skills even with point-and-click cameras on every phone.
Along with these new technologies, new safeguards will need to be developed. And some industries/applications will adapt to the use of the newer tech. Some people will avoid change, some will grudgingly adapt eventually, and others will embrace it from the start. It's always a mixed bag.
What isn't useful is spreading one-sided fear mongering stories that generally are being told and retold by people who do not understand the technology and don't bother to learn. They hear/read BIG SCARY HEADLINE and then form their forever opinion based off that first impression. They then pass on that snippet verbatim to everyone who will listen, and thus no one ever actually takes the time to form educated and carefully considered opinions, but instead gut fear reactions.
For those who wish to actually dig in to some counterpoints:
Neuroscience, Artificial Intelligence, and Our Fears: A Journey of Understanding and Acceptance
Key Facts: Our fear of AI is rooted in the amygdala’s response to uncertainty and potential threats. Fears of AI commonly revolve around the loss of control, privacy, and human value, as AI develops capacities that might outperform human abilities. Addressing these fears responsibly involves understanding that AI mimics but doesn’t possess consciousness, ensuring ethical data handling, and promoting a ‘human-in-the-loop’ concept where AI collaborates with, rather than replaces, humans.
Key Facts:
‘AI Anxiety’ Is on the Rise—Here’s How to Manage It
Does AI Kill Human Creativity or Enhance It? Editorial
@cuddlefaery - I never made the argument about it killing off skills. I don't know where that came from.
@cuddlefaery - Did you even click on my link? It had absolutely nothing to do with A.I. killing human creativity. It's about people using A.I. to make non-consensual deepfake images of somebody.
@Mike403 I wasn't responding to that particular comment/link of yours. Note how I didn't tag you that time?
And I would argue that stating that anyone without Photoshop skills can create deep fakes does lead into the argument about creativity. A lot of artists and those in creative fields are concerned about the future of their careers because of AI generated art. It's a whole tangented debate to the deep fake problem because both stem from the same tech.
@cuddlefaery - My argument was about if someone didn't like you for some reason, they can destroy your life with Deepfake porn. It used to require either being skilled in photoshop or paying somebody to do it. Now with A.I. Deepfakes, it's a more serious concern.
@Mike403 regarding the link you shared and the issue of deep fakes: 1) it's interesting (sarcasm for the topic, not directed at Mike) that for the entire length of the Internet, at least, women have been trying to get revenge porn protection laws in place....and it's only now that BIG SCARY AI is involved that more people are starting to take it seriously. Some jurisdictions have already outlawed revenge porn for several years, but it was only in March of 2022 that it was put into federal law, and it's still hell to try to prove in court even without AI. 2) again, the point stands that revenge porn and fake images existed long before AI. Hell, AI tools for creating deep fakes existed years before OpenAI/ChatGPT brought AI more to the mainstream consciousness. The root of the problem with fake images isn't the tool... It's the system that continues to allow people to get away with using the tool in a harmful manner and that does not properly educate people on safe use of tech. We are not going to fix the system by making people afraid of the tools... especially since adapted versions of the tools are our best option for fighting harmful use.