The Future of AI: People loving Machines
I had a chat the other day about the loneliness epidemic amongst middle-aged men. Who or what can solve it? Potentially AI. What if all those lonely men had somebody to talk to? Who asks the right questions? Who knows their lives and does not judge them?
Remember the movie "Her," where the protagonist, Theodore Twombly, develops a relationship with an AI named Samantha. This fictional tale from 2013 mirrors reality more closely today, thanks to apps like Replika. For $69.99 annually, Replika offers digital companionship – voice calls, healthy habits, coaching, and intimate conversations included. The phenomenon has even reached a point where individuals share their "break-up" stories with their Replika AI companions on YouTube.
And as many of our friendships have moved to ‘WhatsApp only’, why not have a ‘customizable' AI friend? But for this to happen, AI requires personality. However, large AI modules have worked hard to reduce AI personality and maintain a balance between helpful and friendly and neutral (without this moderation, those AI models apparently treat you like your nasty uncle). But let’s imagine a future in which some AI applications begin to develop stronger personalities.
When AI learns to express a spectrum of emotions such as sadness, happiness, fear, anger, surprise, and disgust. Will people then get attached to these machines? It is a possibility. Our minds are programmed to have empathy for beings with feelings and personalities. Just look at how we care for our dogs.
A comparison chart of AI capabilities against human performance in areas such as speech recognition and language understanding shows that AI is already surpassing human performance. If we extend these metrics to include ‘charm’, ‘seduction’, and ‘persuasion’, they all could also start to exceed human performance.
And AI might keep improving indefinitely. As it continues to evolve, it might become more adept at mirroring and responding to our emotional states. It could learn how we react to it and get better at aligning with our needs, values, and worldviews. It could entertain us when we are bored. Cheer us up when we feel down. Talk with us about the things we consider important in the world. It could be funny, sad, or happy. And suddenly, we will find ourselves caring about it.
If you ask me about the biggest threat from AI. Here is the one. We will not let the genie out of the bottle by accident. The genie in the bottle will make us do it all on our own.