June 2025, 7 minutes

Lonesome Doves

Loneliness is, in essence, a subjective thing. As it is commonly understood, we can all feel lonely in a crowd, and, sometimes, when sitting on our sofa, surrounded by nobody, in an empty house, on a quiet street, in a small town, our hearts can overflow with the touch of others. The latter falls more under the umbrella definition of solitude – a typically self-imposed isolation that can often nurture and grow creativity, build a sense of peace, or regenerate the important bits. In other, better words: loneliness is unwanted, solitude sought.

The phrase “epidemic of loneliness” has caught on in recent years, particularly in the post-Covid conversational landscape, but really had its take-off in 2010 with the release of the seminal meta-study on the topic conducted by Holt‑Lunstad, Smith, and Layton. This study found that loneliness and unwanted social isolation was not just bad for one’s mental health, but also had drastic consequences for physical health, too. Taking a given period of time, it was found by the researchers that strong social connections lead to a 50% increased survival rate. This put the health effects of loneliness on par with smoking, and ahead of obesity, physical inactivity and air pollution. The researchers studied three key variables when coming to their conclusions: loneliness, social isolation (the objective state of being alone), and living alone. All three states were found to drastically increase negative health outcomes and mortality. Interestingly, demographics played little role in the outcomes as the same results were found across gender and age, as well as pre-existing health status. The study served as a concrete foundation for understanding something which would hitherto typically be taken for granted, or considered too vague or undefinable to have true measurability or applicability.

A quick search on Google Scholar or Research Gate will result in tens of thousands of cross-disciplinary hits on the topic. A worldwide 2023 study by Gallop found that 23 percent of people reported feeling lonely “a lot of the previous day”. These findings closely echoed a 2022 study by Gallop-Meta which put the number at 24 percent. The phenomena tends to fluctuate with age, following a sort-of Yerkes-Dodson inverted-U curve where it is most prevalent in both adolescence and old age, usually levelling out in the middle years of life. Loneliness prevalence is also impacted by generational affiliation, with members of Gen-Z reporting loneliness at rates of up to a third, compared to the typical quarter found in previous generations.

Undoubtably, it’s a problem. Seemingly a growing one. So, enter solutions. Up to and including technological ones.

Much like getting older, AI is both on the horizon and already here. In rhythm with the advent of the internet itself, followed by the forum decade(ish), and later the social media paradigm we are currently defined by, as much as an Anatolian famer was defined by the Neolithic period in which they dwelt, AI is very much here, and becoming more here with every new instantiation of the technology. So, the question is therefore begged – can AI, a thing which is modelled on humanity itself, help fix this epidemic?

In search of answers, we return once more to the vast academic wildernesses of Google Scholar and Research Gate, where searching for a combination of loneliness and AI (or any bullion word variation therein) will also lead to a boggling amount of hits, include from before the invention of what we now colloquially call AI (that is, the generative ChatGPT iteration, beginning in 2018, but arguably not taking-off until, 2022, with the introduction by OpenAI of the conversational method of interaction), when researchers and academics theorised if “robotics” (a term that now almost feels quaint) and robot companions could be a solution for loneliness and its established negative health results.

The problems associated with the conceptually inhumane nature of AI solving the epidemic seem almost too obvious to mention. But, the etymology of the word mention comes from the Latin to remember so, lest we forget: the lack of genuine interaction seems problematic, to say the least. It’s not a person back there, “talking” to you. It might be a new thing, or may one day become a new thing, that needs a new name and associated bundle of considerations, but it is not a person, the very idea that is in play when discussing the topic of loneliness.

To that end, social skill erosion could (I would argue would) quickly emerge. Talking to people is a skill. And as with all skills, they come easier to some than others, and those who they don’t come as naturally to, will require more practice lest the precious ability weakens. Here we encounter the recursive problem – the over-reliance on AI tools for interaction simulation could slacken real-world abilities, leading to a further reliance on the artificial, the metaphysically false. Now the loneliness is exacerbated, the “solution” quickly becoming a part of the problem.

And what of ethics? Ay, here’s the rub, the “marrow of all”. Is it right, is it good, to make people feel like they are being emotionally understood by something that cannot feel, at least not in the way we conceive of feeling? This question becomes especially pertinent when it is understood that the kind of person engaging with AI to mitigate against feelings of isolation and loneliness is an inherently vulnerable person. In this circumstance, ethical considerations magnify considerably. What of manipulation? Corporate malfeasance? Loneliness in the capitalist age? The motive to data-mine and profit from the epidemic is too much to resist for a righteous robber baron and their legions of minions and, most tragic of all, currently there’s no real mechanism, moral or legal, to effectively stop this, certainly not on a global level.

Naturally, the above enumeration is not exhaustive; just a taster of some of the more obvious problems. The subtler list may be much longer, and would probably contain many second, third, fourth etc, order consequences, which are, due to their sneaky nature, so often the most destructive kind.

As previously mentioned, the era I am referring to is already upon us, like a storm gathering itself above our heads. The Augers have observed this contentious sky, and the birds passing across it, and are ready to deliver the message. The homepage for Replika, the company that appears to be leading the way on this front states it is “the AI companion who cares. Always here to listen and talk. Always on your side”. Does that not set your nerves on edge?

The outrageously cynically named Woebot Health is another Silicone Valley AI startup promising to bridge the gap between tissue & fibre and chips & wires. Founded by a psychologist in 2017, it uses Cognitive Behavioural methods to create chatbots that can help the user through everyday mental health challenges. On their homepage, they state – truthfully, one could argue – that mental health needs have multiplied, whereas support has not. Never let an opportunity go to waste, and all that…

This is not to have a pop at these companies specifically, though I don’t feel like any contemporary corporate entity deserves even a crumb of the benefit of doubt. I’m sure the founders and vested-interest parties, as well as some the users, could convincingly and emotionally argue for the benefits. Though the contrary anecdotal evidence may be just as compelling, depending on the listener. Sweet; good for them. That’s not what this is about. More importantly, I contend, is the necessity of a general (hopefully adult) conversation which society seems either too busy or too disinterested (or, I would argue, too disenfranchised) to have about the proper role of AI in society - how it should be adjudicated, utilised, channelled, even limited. And who should have the final say on such things. As it stands, the word of God appears to lie in the sweaty hands of Tech Bro’s with dubious records on what Aristotle termed the chiefly good.

…

So, is AI the Promised Land? The place that can deliver us as a species from the aloneness of it all? I suspect not. And my impression is that few see it as such. The technological cynicism of this age - the general tiredness that can flirt with nihilism - as well as the very-real, very-material limitations on catabolic energy production and use, mean that the AI Age is starting with a dry, throat-clearing cough, somewhere over there in the corner, where the weird kids hangout. But, despite its whimpering beginnings, it is quickly becoming a subcutaneous layer beneath everything we do, from marketing to manufacturing, medicine to memes. I would suspect that emails written by the human hand and the human hand alone are now a thing of the past. Likewise, the zone of the coder has become flooded with the technology. Additionally, the scope of the arms race between education provider and education receiver has yet to be truly understood.

It was the American philosopher of science and technology, Thomas Kuhn, who in his 1962 groundbreaking work, The Structure of Scientific Revolutions, noticed how science and our collective understanding of the commodious disciplines followed a pattern of stasis and change: paradigm, where theories, methods and standards dominate how scientists interpret data and conduct research; normal science - this is the period in which the science is conducted according to the paradigm; scientific revolution occurs when contradiction in the current paradigm becomes overwhelming and anomalies develop, requiring a new approach; paradigm shift - this is the meaty part of Kuhn’s theory, the revolution of thinking and approach, where old ideas are discarded and new ones taken up; finally we have incommensurability, where the theoretical underpinnings and methodological approaches have so radically shifted, that any comparison with a previous paradigm becomes fruitless, or even impossible. At this point you find yourself in a new paradigm. There you go; feel free to rinse and repeat this process as needed.

It's hard for me to say exactly where in this process we currently find ourselves. One could argue we are somewhere between scientific revolution and paradigm shift, as these new tools are changing both how we interact and how we develop. However, one could also argue with equal vigour that we are in fact still firmly entrenched in the paradigm itself, with recent tech advances merely reifying the current system, and that it offers fundamental contradiction that requires a resolution.

Good point, and good point.

…

The title of this piece wasn’t a throwaway pun, though I believe it would be justified in being so. Rather, the last great American epic is a 900 (depending on edition) page long search for connection in that most lonely and dangerous of places – the frontier.

We all have our own frontier, a personal one, inside us. And it can be a dangerous place, too. And it can be a lonely place. A wild place where all of our time, lived and yet to be lived, merge into a Dalian dreamscape. A place filled with myths and monsters and rushing rivers that seem unbridgeable. But, as with the flawed cattle-pushing heroes of Lonesome Dove, we must forge forward nonetheless.

Naturally, Woodrow Call, Gus McRea and Lorena Wood did not have access to robot buddies or language models that could approximate human interaction. And it’s unlikely their personality types would adapt very well to the possibility. They had no Sonny, no Data, no Rachel. But they had each other, until they didn’t. And then they had others, until they didn’t.

Larry McMurtry can, of course, sum it all up better than this shadow-dwelling author ever could: “It was a fine world, though rich in hardships at times. And if it was not a world that lent itself to happiness, it was at least a world that made loneliness bearable.”

Yes, Larry. It was.

It is.