derived: embodied AI has millennia of precedent

lumnl is built around a counterintuitive idea: objects that contain intelligence have always been part of the human experience. Embodied AI’s must respect and reflect timeless human qualities.

Silicon Valley is obsessed with the unprecedented. We build technologies that unleash the new, new thing. Yet we ignore our deeper roots at our peril. The designer and technologist who understands that our human striving is timeless, that at a foundational level we are building the same things we have always built, can tap into millennia of precedent for inspiration and connection.

I am obsessed with 3 topics that profoundly inform how we should embody AI.

Noogenesis - the cognitive abstraction cascade.

We must design intelligent systems to interact with humans at the appropriate level of abstraction, not blindly chasing intuitive convenience.  Machines must communicate in the medium most appropriate for the message. 

The evolution of intelligence in humans is being mirrored by the expression of intelligence by machines. Noogenesis appears to be on a logarithmic timeline that bends back on itself. It took 200,000 years for human intelligence to shift its leading expression from form → voice → image → text → logic; digital machines have moved back down the abstraction ladder from logic → text → image → voice → form in a single lifetime.

Noticing the pattern of how intelligence evolved in humans and was then built in human machines carries deep implications. This is not merely a fun quirk of history, it is a word of caution and a call to action. The more a device interacts with our primal senses the more vulnerable we are to their manipulations. The more a device interacts with our cognition the more opportunity we have to engage as our better selves. Speech is subconscious, text is conscious. Physical objects are intimate, virtual objects are facile.

We are building angels, let us make sure they address our better natures.

Read the full essay here: lumnl.com/blog/noogensis

Divinity - the qualities humans ascribe to gods & God.

Humans have long interacted with supernatural beings that have qualities similar to those of modern AI systems and software agents.

We must be careful about creating digital gods and even more so digital Gods.  Historically lowercase g- gods were understood to have limited personal knowledge, discrete capabilities, localized presence, and distinct personalities.  Those capabilities continually conglomerated and expended into the monotheistic, Abrahamic uppercase God that lead the world’s cultural understanding of the supernatural divine: omniscience, omnipotence, and omnipresence.

Beware the following characteristics in your nascent AIs, and pay extra attention when omni- prefixes gets tacked on. :

  - personal awareness

  - independent agency

  - uncommon potency

  - ubiquitous presence

Our culture isn’t ready for digital gods, let alone Gods.

Cultural commonality - the unifying threads of human ritual and art.

For a time Structural Anthropology sought true commonality between human cultures. Anthropologists looked at historical records and visited disparate places trying to seek threads that unified the human tapestry, common bonds and practices that seem universal to all humankind.  This became unfashionable in the midst of efforts to reestablish cultural identities for oppressed peoples, to restore cultural precedence and uniqueness for cultures that were displaced by the West’s domineering technohegemony and cultural imperialism.

Today we believe human beings are equal but their cultures are differentiated, yet by looking too closely perhaps we lose the willingness to see sameness in our innate human qualities.

As a designer I want to find places of deep cultural crossover – to seek forms, rituals, interactions that unpin our profound humanness. So I study ancient art forms and rituals, particularly seeking those cultures that practiced refinement and concentration instead of innovation. I study anthropoligists, linguists, and art historians that seek unifying foundations of cultural expression or commonalities.

Next
Next

technosocial constraints on building AI for the real world