AI Observers: The Future of Learning Experiences – and a Prediction for Learning Systems
Thesis: Skill observation with AI-enabled devices will finally make ‘learning-in-the-flow’ a reality.
You are being observed!
Through cameras, microphones, keystrokes—even your body language. In the future of work, AI won’t just assist—it will observe, interpret, and react.
Think about how autonomous vehicles operate today. The Tesla SUV, for example, has more than a dozen sensors constantly scanning its environment, making decisions in real time. That’s not just automation—that’s situational awareness.
Now imagine applying that model to human learning.
Learning by Doing— Your New Skilling Data
Let’s take a real-world example. Say you're running a fast-food restaurant. You hire people on the spot—easy enough. But training them? That’s a bottleneck. Managers get pulled off the line to observe and coach, and the whole onboarding experience is inconsistent at best.
Now insert a wearable AI device. It knows the station, it knows the task, and it knows the sequence. It prompts the learner at the right moment, observes their execution, and adapts. This isn’t science fiction—it’s the logical next step once the right data flows.
This is true “learning in the flow”—not just consuming content mid-task, but being coached, guided, and corrected in real time by an AI observer.
And it’s not just task-based roles. Think of a new manager learning to give feedback, or a team lead practicing communication under pressure. With the right inputs, soft skills become observable too—and coachable.
As Accenture predicts, 20–40% of working hours are poised to be augmented or automated by GenAI in the near future, shifting not just how we work—but how we learn while working. Source
The Hidden Learning System
This leads to a radical prediction: The learning systems of the future won’t be systems in the traditional sense. No dashboards. No checkboxes. No obvious front-end.
Instead, the next generation of learning platforms will operate like invisible copilots—quietly observing, gathering data, and offering assistance exactly when needed. They’ll be embedded in the workflow, not bound to a learning portal.
As Sam Altman recently put it,
“...and people in college use AI like an operating system.”
That’s not just a generational quirk—it’s a preview of how the workforce will expect to learn, adapt, and perform: constantly assisted, seamlessly integrated, and completely personalized. Source
Skills Intelligence: A New Source of Truth
Right now, “skills intelligence” is hot—and for good reason. Companies want to understand what people can do, where they’re growing, and how to close gaps.
But the current approach leans heavily on human trust: manager ratings, peer endorsements, self-assessments, and knowledge recall based tests. These all have value—but they also come with issues such as bias and noise.
AI observers offer customizable constancy, and produce higher value data.
Imagine a system that doesn’t just log completions or accept subjective ratings, but actually tracks a learner’s application of knowledge—memory recall, procedural accuracy, speed, even confidence.
As Josh Bersin notes, the $340 billion corporate learning industry is undergoing a massive disruption toward adaptive, real-time systems that break away from static evaluations. Source
The opportunity here is not just to measure skills—but to witness them.
From Bloom to Behavior
This is where Bloom’s Taxonomy makes a comeback.
We’ve long used it to define learning objectives, mostly through verbs—understand, analyze, apply, create. But what if we built systems that could actually detect those verbs in action?
We’ve simulated this with Likert scales and observation rubrics. But with AI, we’ll need more. We’ll need systems that recognize not just what the skill is, but what it looks like in motion.
That means new data structures:
A framework of skills
A framework of actions
A model that links behavior to proficiency, not just progress
Megan Torrance's work in learning analytics emphasizes this shift—
“Bloom’s verbs are scaffolding for skill development, and designers will increasingly need to pair those verbs with observable behavior and data frameworks.”
(Source: Megan Torrance, Data & Analytics for Instructional Designers) Source
Example Verb Hierarchy
Closing Thought: FitBit for Competence
Back to the Tesla metaphor—what FitBit did for steps and heart rate, AI observers will do for skill development.
And once you’ve got visibility into what people are actually doing—not just claiming to do—you can start delivering learning experiences that feel almost telepathic in timing and relevance.
The AI observer isn’t Big Brother—it’s your new personal tutor. And it won’t ask you what you learned.
It’ll already know.
References
Altman, S. (2024). MIT Fireside Chat on AI
Bersin, J. (2024). The $340 Billion Corporate Learning Industry Is Poised For Disruption. joshbersin.com
Accenture (2023). Reinvention by Numbers. accenture.com
Torrance, M. (2020). Data & Analytics for Instructional Designers.
