Algorithms and Relationships

People use the internet to transmit messages back and forth billions of times a day. In fact you yourself probably text friends and people you know dozens of times if not hundreds in a single day. What if however you were not talking to a person, but the digital ghost of one. What if instead of mourning your grandparent’s death you told them yourself? What if you could continue a romance with someone who dumped you all with the push of a button? Or, perhaps the scariest of all, what if you never knew anyone “real” to begin with? Welcome to the twilight zone of algorithms and ethics that is AI relationships.

This is not, as one might be inclined to think, a full upload of a human mind to a computer. In fact, comparatively speaking on the scale of AI that is available even at the time of writing this article the power of the idea in question is miniscule. Reasonably this is not even AI. So what is it? It is just an algorithm, and a simple one at that. Take my aforementioned example: You are a depressed younger person who recently had a grandparent pass away. You were sad for a few days sure. Then you get the notification on your phone the next morning that the app is ready. You yawn, stretch, and calmly open your phone to see your grandparent’s exact face smiling back at you on the screen. You talk with them for hours as if nothing happened. In fact for years to come you talk to them. Sure maybe less or more depending on your own life, but always with no thought even considering they are or are not alive. They are always there for you, one swipe and message away.

Usually this is the part where I would provide a reality check for the example. Not so with this. The example is precisely that: a reality check. Everything in that rudimentary example is already fully developed technology and economically viable methods of creation. The only caveat I offer is that at the present time I am not aware of any company offering such a service to the public.

So then, how does it work? What is the technology capable of? In laymen’s terms it is an algorithm that you feed data until it talks and behaves like the person it is being based on. You give it all the text messages of that person, their contact data, their recorded audio, their writings, and anything else that has the creative watermark of the person in question. You feed all that data to an algorithm to “train it”. Do that with enough data and then pat yourself on the back: you have brought that person back from the grave in the form of a digital ghost.

There is a variant of this tech on the commercial market. It is different from the concept above in the sense that it is not based on a single human. Instead it is basically an advanced chatbot. You can talk to it just like a person. It can remember things, form memories with you, and develops a personality based on your own. Behold, Replica. Developed and publicly available for years this chatbot technology has advanced far beyond just texting. In fact at the present you can talk to the replica as well as take photos with them and interact with them in Virtual and Augmented Reality.

This technology is unsurprisingly as useful as it is ethically questionable. Fortunately, this does seem to be an issue that is not dangerous in the ways an intelligent digital construct (such as actually transferring a human consciousness to a computer) would be. There is nothing truly “intelligent” about this. It is basically a chatbot trained on a human’s real data to the point of emulation. It is unaware of self and just is providing a familiar output when an input is given. Conversational, sure. Ethically dubious for the bot itself? Hardly. That said there are concerns ethically even if they aren’t directed at the person being replicated digitally. For a simple example of said questions: what are the effects of talking to someone long dead? Will it adversely effect those who use it? What about those who don’t understand the technology such as the elderly or the very young? Is it permissible morally to let a toddler talk to the digital ghost of their grandparents? The answers to these questions are not easily acquired and will be asked more frequently once the technology finally reaches the general public.

Another element entirely that is worth consideration is the sociological effects this technology would have on those who use it to avoid people. They know it is an AI and are comfortable with that fact. In time they avoid any interaction with any normal conscious human for the safety and infinitely customizable algorithm they have grown to know so well. After all, an AI isn’t going to forget secrets you tell it or important dates. It will be better at picking up on verbal cues than any normal person (and even professional) could ever hope to. It will always know what to say and how to say it, and will only get better at doing just that with time. Would people become engulfed in such a relationship, or would would a majority of people still feel it was too artificial? To what scale would this problem emerge?

Yet again despite the potentially troubling scenario where this happens, I believe that most likely we will see the halfway point of this concept become the norm. Assistants that also have personalities will begin to take root in the general public. Your best friend may be a person, but your second best friend will be the AI you have known since you were four. As the technology advances this “assistant” of a best friend may become less a chatbot and more a fully fledged part of your life, much like a secretary or a butler but much more personal.

Regardless of what the future holds though it is as always imperative thought is given to it so that we might shape it for the better. Artificial Intelligence has been advancing at an exponential rate even in the last three years alone. In fact comparing progress from a decade ago to the present would be almost like comparing a Roman chariot to state of the art electric vehicle. The differences are astounding. The tech I have mentioned today will become a consumer product sooner than most would imagine. Until then however we can admire the currently available versions and ponder just what the future of humanity will look like when AI fully arrives.