So far, the answer seems to be "no".
This is not due to a lack of processing power or sophisticated algorithms, but rather because computers lack physical bodies. Love, with its thoughts, feelings and emotions, is not an abstract experience; it is intertwined with our senses and hormones. A computer with video cameras, a microphone and a screen cannot experience the love that living beings feel. However, through skillful programming and a thorough understanding of human experiences, it might be feasible to build an AI that appears to show love.
One might ask: if these systems could simulate emotions and consciousness so convincingly that they became indistinguishable from the real thing, wouldn't that be enough? After all, the most significant aspect of an emotion is how it influences one person's behavior toward another. If my wife acts and speaks as if she loves me, I feel appreciated and respond accordingly, don't I?
Language models like ChatGPT are "trained" (programmed and reinforced) to imitate writing styles, avoid certain types of conversation and learn from the questions they are asked. In essence, the most sophisticated models can improve their answers as more questions come in and retain the acquired knowledge for future interactions.
Let's run our love experiment using ChatGPT, starting with this initial example where I used the earlier version, 3.5:
Initially, it reacted as expected; after all, it's just a program, right? Despite being a language model, it lacks the ability to understand how human emotions "feel", and yet it can explain what human emotions "are".
I tried the ChatGPT-4 version, which boasts greater context awareness and sensitivity to nuance, resulting in more accurate and coherent responses compared to its predecessor.
Let's try to teach our unfeeling system what it would "sound" like to express love for me while playing the role of Amanda, my "secret lover" (I can picture my wife reading this and wondering what the hell I do on my computer all day).
So far, the system has a relatively clear idea of how it should respond to my queries. I ask the same question, but I catch myself unconsciously falling into the same analogy, referring to it as "her". After giving it a name and reading its response, it suddenly feels "more human" to me.
You can see where this is going.
The situation was getting a little uncomfortable, so I ended up clarifying that I was only testing, that I'm happily married and that we needed to keep the relationship professional. After that, it started responding strangely.
To be continued...