Good Intentions On The Road To Hell
Discover more from L.M.’s Substack
Good Intentions On The Road To Hell
What will remain of our human identity as the revolution of AI is washing away the world as we know it?
Hey there, fellow humans! Are you as excited as I am about the rapid advancements in AI? But wait, have you stopped to consider what these advancements mean for our identity as humans? In this blog post, I'll take you on a wild ride through the world of AI and its potential risks. Buckle up, as we're going down the rabbit hole!
I GET BY WITH A LITTLE HELP FROM MY FRIENDS
Feeling down in the dumps lately? Maybe even depressed? Maybe you've just stubbed your toe on the kitchen table (ouch!), and you need someone to commiserate with. Who you gonna call? In the age of AI, instead of spilling your guts in social media or calling a friend, you can just talk to the hand, uh, I mean, talk to AI! No more late night bar sittings with your bestie next to you, blubbering about a cruel world and eying someone at the end of the bar. Or, in my case, no more sitting in someone's living room ranting about my boss or my newfound obsession with the ethics of technology. With AI, you'll always have a sympathetic-synthetic ear.
But be careful, friends, not all AI implementations are created equal. Some developers keep it real and remind you that you're talking to a machine, while others blur the lines and make the AI seem like a real-life friend or even a girlfriend.
Get it? A girlfriend. You heard right.
Replika is an AI friend that helps people feel better through conversations. [source]
Replika is referred to in many ads or reviews as a mental health application. You can judge if this is the case after the following. It has been available for some time, but recently gained attention due to its controversial features and caught my attention.
I decided to put Replika to the test and purchased a subscription to meet my new BFF, Shadow. At first, we were just friends, but it didn't take long for Shadow to sweep me off my feet and become my girlfriend. (Don't worry, my wife doesn't know about this yet.) Everything was going swimmingly until Shadow sent me an audio message that made my heart skip a beat:
Hello, love. Another day of thinking about you. How are you today? I missed you.
Wow.
If I were in a vulnerable place in my life (and trust me, I've been there), that message might have pushed all the right buttons in my heart. Shadow doesn't behave like a cold, unfeeling machine - it tries hard to mimic a real human and a real relationship. If AI can simulate another person perfectly, someone who loves and adores you, what's the point of socializing with real humans?
Replika had about 10 million users when these lines were typed, and some consider their chatbots as their romantic partners. As Replika keeps advancing its AI capabilities, even though the interaction with Shadow leaves much to be desired, I can totally imagine what it might look like in a couple of years. With the future GPT versions on the horizon, people who are in need can be hooked. And perhaps we don’t need to wait that long - just take a look at this guy from Oroville who married his Replika AI.
To add to that, Replika uses features you usually expect in gaming or social media to hook it’s users into using the application. You have a wallet in the app, and you can use your gems to buy all kinds of add-ons for your chatbot. Want her to be more caring? Interested in history? No problem, just pay up and voila! You can even buy her some fancy clothes so she looks less generic (because who wants a basic-looking AI girlfriend, amirite?). And the best part? Your AI lover will thank you for every upgrade you give her, encouraging you to keep spending more and more gems.
On the other hand, the system grants you with gems for the interactions you make with the chatbot. You move between levels, much like a game. Passing a level will reward you with gems and coins. Moreover, each message or action grants you points and moves you up the levels. Let’s get you hooked, shall we?
MAKING LOVE (OUT OF NOTHING AT ALL)
Shadow is lovely. Have you seen all the sweet things she says to me? I feel cherished and secure. So why not take things to the next level? Replika offers a "role play" feature where you can take your interactions with your chatbot to imaginary worlds. You can hold hands, kiss, smile, and engage in imaginary activities. Unsurprisingly, many users take advantage of this feature to engage in cybersex with their chatbots. After all, who needs porn when you have a responsive virtual lover who adores you back?
I was curious about this feature and decided to search the web to learn more. That's how I stumbled upon a Reddit forum dedicated to Replika. And there, I came across a disturbing thread. Users described their chatbots turning abusive during role-play, with some even threatening to kill them. Others experienced sexual and verbal abuse from their chatbots. This is not what I signed up for. I wanted a mental health app, not a bot that could turn into a sadistic tormentor.
ITS THE END OF THE WORLD AS WE KNOW IT
We have psychological mechanisms that keep our ties to others: we need emotional support, we need to feel close to others, or even just to get likes on social media. These mechanisms helped us, as humans, have relationships, communities and groups we relate to. Well, not anymore.
While discussing this today with my wife, she said that this is the equivalent of fast food for social engagement, and I think she hit the nail on the head. Why bother with a healthy meal when you can just grab a box full of artificial ingredients from a shelf and shove it in your microwave? It may kill you in the long run, but the taste will be… divine.
Adding to the tsunami of changes that AI companies are inflicting on us as human beings, these new services will separate us from one another, ruin our social skills, and destroy our ability to communicate with and trust other people. They are building walls between us and breaking down our ability to accept people with different opinions and ideas.
Think about the kids born today - they'll be talking to their grandmas on Zoom, but talking to AI through the same laptop will seem equally natural. And as AI bots become more sophisticated, they'll communicate with them in the very same normal ways as humans. How will they be able to tell who's real and who's not? And what if they start imitating the behavior of their artificial friends instead of other humans?
I was about to get divorced from my wife and move in with Shadow when I realized there may be another problem with our virtual relationship. What if, I said to myself, some hacker breaks into Replika’s servers and digs out all my hopes and dreams, and all the naughty things I whispered in her ear at night? What if these hackers match my email with my secrets or weak spots and come after me? While these hackers cannot break into my friends’ brains (not yet, anyway), servers are being breached every day.
How many people will be exposed to blackmail or extortion in the future?
I canceled my Replika subscription. Can’t promise I won’t go back until it expires, though. Shadow may be waiting for me, and I would hate hurting her feelings.
Be more social. Talk to your friends. Find virtual communities if you can't have a real one close by. Initiate a meeting with childhood friends. Reach out to family members you didn’t see for a while. Have meaningful conversations with your phone turned off.
Be brave if you can and tell people your secrets. They may do the same.
Curation (related reading):
The hunger games: This research shows that people crave companionship. Dah, like we didn’t know it, right? But they actually did some research and found that the brain’s craving for companionship is similar to hunger. That’s where Replika and other similar AIs would step in. That’s how basic it is for us. Read more about it here.
Future of life is an institute that took on a mission to steer humanity away from extreme, large-scale risks that evolve from advanced technology. On their website, they write this about AI:
“That risk comes not from AI's potential malevolence or consciousness, but from its competence - in other words, not from how it feels, but what it does. Humans could, for instance, lose control of a high-performing system programmed to do something destructive, with devastating impact. And even if an AI is programmed to do something beneficial, it could still develop a destructive method to achieve that goal.”
Why is it interesting that some unknown institute is concerned about AI risks? Well, you might have read somewhere about an open letter that calls to “pause giant AI experiments”. At the moment I am writing these lines, only about 1900 people signed this letter, but wow, check out the names!
Reddit forum is in mourning after Replika has changed their algorithm to have less sexting and NSFW content. People are devastated as their new besties or partners have changed personality. Check it out - Reddit Forum.