Artificial Intimacy News #6
A field report from the place where 21st Century technology meets human behaviour, culture, and evolution.
A quick thank you to those of you who subscribe. This week, Natural History of the Future finally crossed into three-figure subscription. According to the Substack lore, the first 100 subscribers are the hardest, and I eagerly await exponential growth. Whether that eventuates or not, I would like to thank you all for subscribing, following, but most of all for reading.
In this issue:
The need for proper evidence on whether virtual friends can help with loneliness and mental health.
A substantial and far-reaching perspective paper on conversational AI.
More (and more) on AI and chatbot romance.
Futurist
picks out some parallels between virtual friends and pornography, with thoughts about mechanisms like neurotransmitter balance.Art: an installation. And my thoughts on a possible AI-abetted creative explosion.
Podcast of a panel discussion on the dangers of “counterfeit people”.
Why I claim that humanity’s much-touted demise is the least interesting thing about the rise of AI.
Insight: Do virtual friends ease your loneliness or improve your mental health? Show me the evidence!
Virtual friend platforms, or at least those humans who market them, are fond of pointing out the benefits of their products: easing loneliness, and helping people cope with anxiety, depression, suicidal ideation. You might be surprised to learn that the evidence for such claims is rather scant.
Some evidence is beginning to trickle in. First, inadvertent natural experiments like the one Replika.ai created when they shut off some important capabilities in early 2024. Many users experienced genuine distress, even grief, and reported heightened anxiety. Turns out Replika’s hype wasn’t without some substance.
A study out of the U.S.A. surveyed 1006 student Replika users, finding they used it in a variety of ways. Notably, the students reported that Replika diminished their loneliness and 3 per cent said that using Replica halted their suicidal ideation.
By contrast, reearch on how conversational agents work on human minds, building closeness and a sense of the agent being person-like, is starting to power up. See the next item in this newsletter which discusses a recent review. The public funding environment for suitably disinterested research on these technologies is currently weak because the proposals fall between traditional academic disciplines. But the research community is building.
When it comes to user well-being, however, neither “lucky” experiments nor user surveys constitute acceptable levels evidence. The claims made by chatbot companies – including therapy chatbot companies – demand clinical-standard evidence.
Proper randomised controlled trials might be slow and expensive, but we expect that kind of evidence from drug companies. We have a right to expect if from any organisation that makes claims about the mental health benefits of their products.
Tech companies like to move fast and hope that if they break things, those things aren’t precious. “Like people’s children”, some might mutter.
Headlines & Highlights
Perspective article on the pros and cons of human-like conversational agents
Technologies that can talk with users are enjoying quite a moment. They tap into a very human tendency to deal with other things using the mental tricks and tools we use when dealing with other people. This tendency to anthropomorphize is especially strong toward things that talk to us, or that we think are talking to us. Safe to say that today’s conversational technologies get the full anthropomorphic red carpet.
The prestigious journal Proceedings of the National Academy of Sciences just published a perspectives paper by Sandra Peter, Kai Riemer and Jevin D. West that considers how human-like conversational technologies have become. They review the fast-accumulating evidence for both benefits and downsides of anthropomorphic agents, and consider what should be done.
Not everyone has time to read the whole paper right away. Peter and colleagues also published a shorter article at The Conversation which the editors (who decide on titles at TC) gave the rather breathless title “Evidence shows AI systems are already too much like humans. Will that be a problem?”.
We are only one AI chatbot away from falling in love
Debra Soh gets on the Artificial Intimacy party bus and writes for Toronto’s Globe & Mail.
Like any technology, today’s AI humanoids will only grow more sophisticated with time. A demand exists for them; they will not devolve. We should ask ourselves what their popularity reflects in us. Having successfully trampled over the Turing test, they grin triumphantly, awaiting our response.
The Psychology Behind AI Romance
Some Substack insights from
on Artificial Intimacy.Oxytocin, Pornography, AI companions, and pair-bonding
Over at his excellent Gen X Science Fiction & Futurism Substack
goes deep on some connections that are often hinted at, but science has yet to fully catch up.There’s no official line of demarcation between collapse and transformation, but if future civilizations bother to study our remains, one signature may stand out: The moment humans stopped learning to live with one another— And started optimizing to be alone.
BONUS: I was delighted to stumble across
’s Substacks (plural) because I still recall a really fun interview we did when Artificial Intimacy first came out. If you’re a podcast listener, then take a listen.The many ways AI is being used on dating sites
Megan Wallace writes at Dazed:
In the main, discussions about AI and dating often focus on platforms like Replika and Anima, which offer humans the chance to find (train?) the AI companion of their dreams. But, it turns out, the digisexual revolution starts with us first – and AI images are only the beginning. In fact, what appears to be even more popular is the use of tools like ChatGPT in order to craft bios, replies and even streamline dating admin.
Art and Exhibitions
Cyber Deities and Digital Pastoral: Reconstructing Faith and Nature in the Artificial Landscapes of the Future
Mark Westall writes in FAD Magazine about Bobby Zhaocheng Xiong’s digital installations, including New Deity (below).
My Media Appearances & Posts
Do Artists Have a Future? Getting from Guernica to GPT
I had unfinished business with Pablo Picasso. That’s what I told my travelling companion when I explained that we simply had to spend a recent Friday morning at Madrid’s Museo Nacional Centro de Arte Reina Sofía. Specifically, seeing Picasso’s Guernica
Podcast: Counterfeit People (2024)
Panel at 2024 Festival of Dangerous Ideas, Sydney featuring Lizzie O’Shea, Patrick Stokes, Emily van der Nagel, & me in the chair
The late philosopher and scientist, Daniel Dennett talked about ‘counterfeit people’ as one of the great dangers of AI – but are we now willing to court the same dangers through our adoption of multiple identities. Moving from the confinement of physical reality to the landscape of the metaverse, where looks, preferences, and genders are limitless, we can each acquire many digital selves. Is a ‘virtual you’ a truer reflection of your deepest self – revealing desires and aspects that otherwise remain hidden? What is the human cost of leaving the physical world behind? What are the ethical implications of living without boundaries in a digital space where the borders between morality and legality, in the ‘real world’, begin to dissolve?
Humanity’s demise is the least interesting thing about the rise of AI
“I am agnostic about the end of humanity.”
At least that’s what I tell people at parties who ask about what I do. Of all the things I work on, my attempts to discern how artificial intelligence might influence human evolution are the most translatable to standing around drink-in-hand chatting to new acquaintances.
Your ancestors were inbreeders. So were mine.
When writing about evolution, writers often begin by reminding our reader that they are an evolutionary success. That they come from an unbroken lineage of successful ancestors. How could it be otherwise? You had a mother and a father. At least biologically, you did. Your parents each had a mother and father of their own. And that same mundane pattern .....
Republished from other platforms … for reasons
Male sexual despots rewrite history
“Heredity”, opined the pioneering cultural anthropologist Alfred L. Kroeber in 1915, “cannot be allowed to have acted any part in history”. I have yet to encounter a crisper expression of the view that biological explanations have no place in the study of societ…
Credit where credit is due, it was the Padverb Podcast that put us together:
https://open.spotify.com/episode/4BZ7iIItEDOBA0u2yb1v5z
Another great installment, Rob!