1. What is AI Fluency? A Broader, Human-Centered Definition

‍

The 4D Framework, developed by Feller and his collaborator Rick Dakan, identifies four core competencies for working with AI.

‍

“We observed that AI was creating brand new ways of working. It wasn’t just incremental innovation, it was disruptive,” said Feller. “And when you create new ways of working, you create new roles. That means we need new skills, new competencies, new knowledge.”

‍

Each “D” maps to a distinct area of competency:

• Delegation: What work should be done by AI vs. humans?
• Description: Can we clearly articulate what we want AI to do?
• Discernment: Are we able to evaluate AI outputs with care?
• Diligence: Are we acting ethically and responsibly?

‍

Feller emphasized that these aren’t steps in a process; they’re overlapping skill zones for individuals and teams to develop.

‍

2. The Framework Offers Shared Language, and Restores Human Agency

‍

One of the biggest strengths of the 4D model is that it helps organizations speak a common language about AI use—especially in times of change. Feller emphasized that one of the biggest barriers to building AI capability inside organizations is lack of shared vocabulary.

‍

“Just start using the language,” Feller said. “It focuses attention away from the tech and back onto the human.”

‍

He also emphasized the empowering nature of fluency:

‍

“I often say to people, accountability is our agency. We're worried about replacement by AI, and that narrative is concerning. But this framework opens up new ways of working. It restores our ability to shape how AI is used.”

‍

Tip for learning businesses: Equip your clients with a shared, simple model (like the 4Ds) to guide discussions and design training experiences. It helps teams make AI thinkable and talkable, which is the first step to safe experimentation.

‍

3. Skillful AI Use = Human-Centered Thinking

‍

While many teams focus on cost cutting through automation and efficiency, Feller and Faktor argue that the real opportunity lies in value creation through augmentation and transformation. Feller made a key distinction between the technical skill of prompt engineering and the deeper fluency needed to make AI useful across roles.

‍

“Description and Discernment are what we think about when we think about AI literacy skills, prompting, engineering context. But Delegation and Diligence are just as important. They both help us answer  ‘why’ and ‘why not’.”

‍

Candice echoed this in her own application of the framework inside Disco:

‍

“Discernment is a real lived experience for us. We’re constantly evaluating the behavior of AI systems: Are there hallucinations? Are the outputs useful? That shift, where workers become editors rather than just creators, is real.”

‍

Design cue for trainers: Move beyond cost-cutting stories. Start designing learning moments that empower workers to collaborate with AI in creative, cross-functional ways. This drives both engagement and impact.

‍

4. AI Fluency Isn’t One Size Fits All – Context Matters

‍

One of the key insights from Feller and Dakan’s work is the need to localize the AI Fluency Framework  to different contexts, whether academic disciplines, job functions, or industry use cases. When it comes to embedding AI use responsibly, Feller and Dakan suggest starting with Diligence, not Delegation.

‍

“Let’s just talk about where the boundaries are we want to set for ourselves,” Feller said. “What does success mean for us in terms of ethical commitment, whatever those values are?”

‍

Feller stressed that understanding the nature of the technology is critical to making good decisions. This is especially critical in environments where hallucinations or bias can surface.

‍

“If you think about something like a language model as being essentially a database, this is not going to go well for you,” Feller warned. “You're going to misdelegate. You're going to misunderstand what kind of work it's best at.”

‍

For training organizations: Build in space for learners to adapt the 4Ds to their specific environments. Use role-based fluency rubrics, real scenarios, and peer collaboration to drive depth.

5. Training Design Must Be Contextual and Experimental

‍

The best way to build fluency? Create space for experimentation, reflection, and conversation.

‍

“We need to fairly rapidly start to de-stigmatize the AI conversation,” Feller said. “Pick a framework. Give your people shared language. And make time—actual time—for teams to talk about what they’re doing and learning with AI.”

‍

His call to action for learning leaders:

• Let teams “localize” the framework to their roles and environments
• Capture and scale internal best practices
• Move past fear and control toward guided experimentation and constant knowledge sharing

‍

Starter move: Host a fluency workshop using the 4D Framework to baseline your learners. Get them talking about how they already use AI, and where they want to go next. Use Disco to turn that into a lightweight, high-impact cohort experience.

‍

Bonus: Free Courses Backed by Anthropic

‍

Feller and Dakan worked with Anthropic (with support from the Irish Higher Education Authority) to create open access courses built around the 4D Framework:

• AI Fluency: Framework & Foundations https://anthropic.skilljar.com/ai-fluency-framework-foundations
• Teaching AI Fluency https://anthropic.skilljar.com/teaching-ai-fluency
• AI Fluency for Educators https://anthropic.skilljar.com/ai-fluency-for-educators
• AI Fluency for Students https://anthropic.skilljar.com/ai-fluency-for-students

‍

“Being able to articulate and communicate, that comes from a position of subject matter expertise,” Feller explained. “When you combine that with AI, you create emergent value, not just automation.”

‍

All of these are available under a Creative Commons license and designed to help anyone, from instructors to students to corporate teams, build foundational fluency in how to think, act, and collaborate with AI responsibly. Links to all courses and other resources can be found at http://aifluencyframework.org

‍

Why This Matters for Learning Businesses

‍

Disco is built for the builders of learning – training organizations, learning designers, and education entrepreneurs. As Candice noted in the conversation:

‍

“We’re not just here to make learning with AI easier. We’re here to help people learn how to work with AI, better.”

‍

If you're building training programs, you have an opportunity to help clients move past generic awareness and into true capability, grounded in responsible, creative, and collaborative use.

‍

The 4D Framework gives you a starting point.

‍