The Gestalt of AI
I’m just scratching the surface, but I consider The Gestalt of AI a very interesting topic to investigate and explore.
While listening to Empire of AI by Karen Hao, which I highly recommend, I find one thing very interesting and mind boggling in particular: the Anthropomorphization of AI. Bottom line: we try to replicate human intelligence and sensory capabilities. And we develop human-like relationships with AI. This debate is going on since decades, not just since the generative AI or AGI hype. What was rather hidden in research labs in the past has now become a public discussion.
Apparently, humans need a Gestalt to deal with. Something more tangible than a pervasive and incredibly abstract „black box“. But does this Gestalt need to be human(oid)? Or do we just lack imagination and culture to explore beyond this notion, or beyond science fiction clichés?
While researching devices for clinical treatment (e.g. dementia or pathological loneliness), I came across PARO and ElliQ. They are both tangible, non-humanoid robot Gestalts to put in the hands of patients, but they both have a very different approach to them. One replicating the traits of a baby harp seal as closely as possible. The other one a more „machine-like“ appearance that leans towards human behavior patterns by gesturing and pulsating light with its lamp-like „arm“.
PARO is therapeutically successful since more than a decade, but commercially challenging due to its high cost barrier. ElliQ has strong commercial growth thanks to consumer-friendly pricing but lacks therapeutic evidence. Both gave me the same strange feeling when initially looking at them.
Both made me think more and dive deeper into investigating and exploring Gestalt of AI. How can we develop new, post-anthropomorphous languages, metaphors and rituals?
PS: The term Gestalt of AI came to my mind kind of randomly, but when I googled it, I found this academic essay called The Gestalt of AI: Beyond the Holism-Atomism Divide published by Hannes Bajohr in 2021.
