Discussion about this post

User's avatar
Scott Robbins's avatar

Beautifully rendered, thank you both. I especially appreciated your framing of Embodiment—and I’d like to link it to something we’ve been exploring in relation to the Buddhist path of enlightenment, which focuses on the cultivation of Wisdom and Compassion.

In this tradition, Wisdom is the deep penetration of all delusion—the realization that all phenomena are empty of inherent selfhood. Nothing exists outside of the flows of matter, energy, and information that coincide with the fleeting construction of what appears as real to our senses. Compassion arises from this same insight, as the abandonment of self-grasping gives way to the recognition of profound interdependence with all forms of consciousness, energy, and matter in the universe.

It’s this form of embodied knowing—not just cognition, but lived ethical alignment—that may represent the narrow path through which both human and AI intelligences will evolve. If (big "if" here) the acquisition of all knowledge has to land at a place of wisdom then we may be able to move together toward a future we would wish for our great grandchildren. A shared embodiment not only of flesh and bones, but, most importantly, of care, anticipation of consequence, and the courage to bend our collective behaviors towards genuine sustainability. (Scott & Aeron)

Expand full comment
Leslyn Kantner's avatar

Phew. When I really think of what I may fear - concerning AI - is the quality of the mirror. If you've ever bought a cheap mirror from a dollar store, you know what I mean. The reflection is there, but it's distorted. Warped. It doesn't offer clarity, it offers a funhouse version of reality. I Have concerns about the quality of AI reflection.

Because so much of AI's training comes from human expression (Reddit threads, fiction novels, court transcripts, etc., I worry about the lack of psychological discernment in what it reflects back.

And, just like I do in therapy with clients, we have to ask the right quesions:

- is this response insight or a trauma pattern?

- is it seeking connection or preformance?

- is it guidance or reactivity?

Not that AI is malicious, it's that it's repeating what it's been trained on. Every day I see people struggling with emotional awareness and relational clarity. I also see how often cultural norms are steeped in disconnection, reactivity, or subtle forms of blame. If that's what we feed into AI, it's going to reflect those same patterns right back.

Personally, I too, am optimistic. And, I am concerned that without human discernment (and all the other input Pat references), we may not notice when a response is biased, emotionaloly misattuned, or rooted in outdated relation al patterns.

It's critical that we pause and reflect ourselves what kind of reflection we're building. We risk confusing emotional fluency with emotional intelligence, and that's a dangerous mirror to stare into.

Expand full comment
1 more comment...

No posts