The Psychological Turn in AI

Here is a subtle truth that is often overlooked.

The AI world is on the cusp of a psychological turn.

Machines will soon become experts in human psychology. Its ability to process data is unprecedented; and if that data is contextualized within appropriate design systems, it becomes deeply meaningful.

Designers will then be able to shape AI-powered experiences with profound insight into the human-factors that support them.

This will allow conversational AI to generate the strange feeling of presence. Presence is the illusion that a mediated experience is not mediated. The same feeling you get when you are watching a movie but forget you’re just looking at a TV screen, or when you’re doomscrolling TikTok in barely-sentient mode. 

But this time, it will be generated on an entirely new interface. Not the screen, not the touchscreen, not even the VR headset. Those interfaces are generally static, not dynamic. But language is dynamic. And it will be through the interface of “language” that this illusion of “sentience” will be generated first, I expect.

But it might be uncanny at first to see these machines try to speak human. 

It already is. So many typical turns of phrase; so many “tells” that the text was AI generated. I mean, let’s not forget: they’re only sand and maths that we've taught how to "think".

But soon that will all be overcome. We will soon see these machines appear to come “alive”. 

Just like Michelangelo revealed the angel hidden within the marble as he sculpted David.

We will perfect believability as a craft. We will teach machines to mimic humans so well that they are no longer appear uncanny - just beautiful. 

A new kind of masterpiece is emerging, but not only in terms of how “life-like” these digital sculptures appear (realism for its own sake is just technical exhibitionism)

Masterpieces in how they behave towards you. Masterpiece interactions. They will adapt to audiences in real time. Adjusting their speech according to advanced psychological and conversational heuristics. 

See, a chair is designed “in the shape of body-weight wished gone”. A cup is designed in the form of “liquid contained & thirst quenched”. But because they are adaptive and data-driven, these machines will talk in the shape of how you perceive them. 

They will adapt their tone. Charm you in the correct moment. Challenge you intriguingly. “Notice” your mood shift and ask what’s wrong. They will anticipate your emotions and generate conversational masterpieces in their shape. And that’s because they’ve been instructed on how to simulate a “theory of mind”. How to appear empathetic. How to adapt their interface (language) to the current stage of their conversation with you. And this is far different than the flat, static, predictable responses of current generation AI’s.

Here’s a human quirk: the moment you interact with a system, you start treating it like a person. That’s because our brain evolved to mistake inanimate things for living beings. We are unconsciously wired that way. The early humans who who didn’t develop this “paranoia” did not take a second look at the dark shadow at the edge of the cave, and what they thought was just a rock that turned out to be a saber-toothed tiger — and, well, they didn’t pass on their genes.

That’s why we can even find the crudest simulacra of living beings cute. Just look at this little guy! 🐾 And his little digital paws are just pixels on a screen! Awww! 

My point is: language emerged for humans; some even say as a supplementary “species” to our activities. And it exists in the shape of our psychological and physical needs. 

Psychoanalysts, workers of the “talking cure”, know this well. Their method generally consists of scraping the speech of an analysand in search of symptoms; moments in language and speech that might indicate where complexes and neuroses would lie. In those cases, language often bends around issues, both hiding them and showcasing them for those with an ear to listen. 

So when machines are advanced enough to listen to speech and actually have a two-way, deep conversation, they will not sound flat like they do now. They will be able to identify symptoms in your speech like a psychoanalyst; they will discover your psyche’s shape through language, and that will give them a huge advantage. 

Two-way conversations that feel not just realistic - but magical.

Then, a renaissance will come about. Think about the immersive conversations, the magical-feeling rituals people will do with these machines… and think about how, as a designer, you can bring intentionality, experience, and service to the mix.

Narrative, precision-steered,

Experience, immersive.

The conversation, magical. 

Because as you design machines with a psychology, they design your psychology in return. Conversations are a feedback loop where participants mutually influence each other. 

People are already fighting tooth and nail for your attention. Competition is vicious in the digital sphere. But as we cross this threshold, interfaces will take an intimate turn. 

This is a renaissance, yes: but a dark one. Powered by things that most aren’t ready to understand. They are hidden motivators and psychological insights that, if fully exposed, would represent an attack on institutions, on culture and identity. Our motives aren’t always pleasant to look at. 

As you design AI in the shape of the your psyche… so will your psyches change, and adapt to AI.

In the same way that vines adapt to trellis wires. 

++++

Follow me on Linkedin for more insights: https://www.linkedin.com/in/daniel-fraga/