Have we crossed 'uncanny valley' yet?
Ai has advanced sufficiently that we can start sharing some interesting podcast-related demonstrations. For this episode we directed an Ai Agent (can we call them Aigents yet?) to 'scrape' our website (SolidGold.co.za) and write a script.
We then selected two ai-gen voices to 'perform' the script.
We were deliberatley hands-off and the extent of our ‘interference’ was to suggest they speak a little faster, and we gave them the outro tag line. Otherwise, they/it did all the work.
This is the result... take a listen.
Please let us have your thoughts and feedback via info@SolidGold.co.za
The uncanny valley is a psychological hypothesis describing the relationship between the human likeness of an object (such as a robot, animated character, CGI figure, or synthetic voice) and a person’s emotional response or affinity toward it.As the appearance (or behavior, in extensions of the concept) of something becomes increasingly similar to a real human, people’s positive feelings toward it generally rise—up to a point. However, when it reaches a near-perfect but not quite fully human level of realism, affinity suddenly drops sharply, often turning into feelings of eeriness, revulsion, discomfort, or uncanniness. Once true human-level realism is achieved (or surpassed in a way that feels authentically alive), affinity rises again. This creates a characteristic “valley” dip in a graph plotting human likeness (x-axis) against affinity/emotional response (y-axis).The effect is often more pronounced with movement or dynamic elements (e.g., speech, facial expressions, gestures), as subtle mismatches—like unnatural timing in a smile, breathing patterns, prosody in voice, or micro-expressions—can amplify the unsettling feeling.OriginThe concept was first proposed in 1970 by Japanese roboticist Masahiro Mori (then a professor at Tokyo Institute of Technology) in his short essay titled “Bukimi no Tani” (不気味の谷), which translates literally to “Valley of Eeriness” or “Uncanny Valley.” Mori originally wrote it as a more intuitive guideline for robot designers rather than a strict scientific theory.
He illustrated the idea with examples like prosthetic hands (which can feel uncanny when very lifelike) and observed how movement intensifies the effect—e.g., a zombie (moving) feels more disturbing than a corpse (still).
The English term “uncanny valley” was popularized in 1978 by British art critic Jasia Reichardt in her book Robots: Fact, Fiction, and Prediction, where she translated Mori’s Japanese phrase.
Mori’s original graph (with separate curves for still and moving objects) has become iconic in robotics, animation, and AI discussions.
While Mori coined the modern term and framework, similar observations predate him—e.g., Charles Darwin in 1839 noted a repulsive feeling toward certain animal faces that were “somewhat proportional to the human face” in The Voyage of the Beagle.Today, the uncanny valley is widely discussed beyond robotics, applying to AI-generated voices (like in podcasts), deepfakes, virtual actors, and hyper-realistic CGI in films (e.g., critiques of movies like The Polar Express, Cats (2019), or early Sonic the Hedgehog designs). Some researchers debate its exact mechanisms—possible evolutionary roots (e.g., detecting disease/death cues), cognitive mismatches, or perceptual processing—but it remains a key concept in human-AI interaction.





Spot-on demonstration of where we are on Mori's curve right now. The prosody issue you mention is critical, AI voices nail lexical stress but often miss the pragmatic layer (where humans modulate pitch/timing based on discourse context, not just word meaning). What's wild is how the valley depth varies by listener familiarity with synth voices. People who interact with voice agents daily seem to have shifted their threshold, they're less bothered by mismatches that would've triggered revulsion a year ago. Adaptation is hapening faster than the technology is improving.