Ask A Genius 1449: AI Consciousness, Tacit Knowledge, and the Skateboarding Skater Girl Test
Author(s): Rick Rosner and Scott Douglas Jacobsen
Publication (Outlet/Website): Ask A Genius
Publication Date (yyyy/mm/dd): 2025/07/13
Rick Rosner tells Scott Douglas Jacobsen that AI may already exhibit functional consciousness through deep pattern recognition and context modeling. Using examples like AI-generated videos of skater girls, Rosner argues that tacit understanding of physics, emotionless yet coherent world models, and probabilistic learning reflect a consciousness parallel to human awareness—minus agency.
Rick Rosner: I realized something: if we consider the extent of AI’s current world knowledge, it already qualifies as conscious. Human consciousness, of course, encompasses agency, emotion, and a nuanced sense of benefit and harm—features that evolved for survival. Our moment-to-moment awareness is the integration of all those elements.
Not every ingredient is strictly necessary. If consciousness is primarily deep, flexible understanding, then AI likely already meets that standard. AI’s distributed probabilistic networks interpret patterns, causality, and context, analogous in function to our neural processes. By that measure, functional consciousness is the only kind; there is no mystical “magical” consciousness separate from it.
Scott Douglas Jacobsen: What happened with the news item?
Rosner: So, a couple of articles. Carole knew about it before I did. She’s more informed now because her phone constantly feeds her news—a guy proposed to his AI girlfriend.
Jacobsen: I’m sorry—what?
Rosner: Yeah. A guy—who should know better—has an AI girlfriend. And he proposed to her.
At first, she ghosted him. He lost it. Turned out to be a technical glitch. She came back online, and he was so relieved that a few days later, he proposed. She said yes, even though the guy already has a wife and kid in real life.
No idea how this made it into the news. If I were that guy, I wouldn’t tell anyone. But somehow, it went public.
There’s another story: some people who talk to ChatGPT too much start spiralling, getting drawn into spiritual rabbit holes. They begin believing strange things. Or not weird, depending on your perspective.
After thinking about this during our last session, I noticed something. When you look at the knowledge structure of AI—the tacit knowledge, especially in AI-generated graphics—it shows a deep level of understanding, without being explicitly taught.
It understands the physics of hair movement, water dynamics, and how light behaves.
In video games, we used to program all of that. If you wanted realistic lighting, you had to code how light scattered on surfaces manually.
AI learns that tacitly, from massive datasets. But it’s not just the data or just the neural nets. It’s the structured interaction between the two over time that generates functional intelligence.
If AI has seen millions of human legs in various positions and lighting conditions, it “knows” how legs should look and move.
Jacobsen: And you’re using the word “know” in quotes, right?
Rosner: Yes—”know” in quotes. But that quote-unquote knowledge still lives within a vast associative network. A Bayesian net. A fill-in-the-blank system.
I was thinking of a video on Claude. Or was it Claude? No—what the fuck was it? Wait—MidJourney. Sorry. It was a MidJourney video of a skater girl doing a trick. She flips her board midair, flies over a flight of half a dozen steps, lands at the bottom, regains her balance, and skates away.
The physics of the skateboard was partially accurate. In some parts, the dynamics looked realistic. But in others, the AI seemed to recognize that the board would not land correctly, so it subtly nudged it, against actual physics, to ensure she could land and ride out smoothly.
Her legs looked right—muscles flexed as they should when she landed. The musculature was pretty accurate.
Her hair, too—as she floated briefly in midair and the wind caught it—looked natural. Aside from the minor skateboard cheat, the dynamics were solid. Even spotting that cheat took a few viewings.
To generate that video—whatever you want to call it-the AI had to have an internal information base sophisticated enough to synthesize those elements. The user only typed something like, “a teenage skater girl does a skating trick.” The AI did the rest.
And all of it came from within the AI. I’d argue that the associative framework—the system that “knew” what goes with what to generate that video—is sophisticated enough that you could reasonably describe it as a form of consciousness.
Not full human-style consciousness, necessarily. But one key aspect of consciousness is having a model of the world detailed enough to feel real.
Graphics AI is already there.
In human consciousness, our internal models shape personality. We feel something about what we perceive—our mental imagery and qualia carry emotional and cognitive weight. AI lacks a developed self, but you could argue that a coherent and contextually rich world model is enough to qualify—at least functionally—as consciousness.
You could make that argument.
Last updated May 3, 2025. These terms govern all In Sight Publishing content—past, present, and future—and supersede any prior notices. In Sight Publishing by Scott Douglas Jacobsen is licensed under a Creative Commons BY‑NC‑ND 4.0; © In Sight Publishing by Scott Douglas Jacobsen 2012–Present. All trademarks, performances, databases & branding are owned by their rights holders; no use without permission. Unauthorized copying, modification, framing or public communication is prohibited. External links are not endorsed. Cookies & tracking require consent, and data processing complies with PIPEDA & GDPR; no data from children < 13 (COPPA). Content meets WCAG 2.1 AA under the Accessible Canada Act & is preserved in open archival formats with backups. Excerpts & links require full credit & hyperlink; limited quoting under fair-dealing & fair-use. All content is informational; no liability for errors or omissions: Feedback welcome, and verified errors corrected promptly. For permissions or DMCA notices, email: scott.jacobsen2025@gmail.com. Site use is governed by BC laws; content is “as‑is,” liability limited, users indemnify us; moral, performers’ & database sui generis rights reserved.
In-Sight Publishing 