A competency model for effective, efficient, ethical, and safe AI collaboration — and how it might strengthen our existing approach.
Use arrow keys or buttons to navigate →
Having access to powerful AI doesn't automatically mean we know how to make the most of it — or engage with it responsibly. The 4D Framework gives us a lasting set of competencies that remain relevant even as AI evolves.
AI Fluency is defined as the ability to interact with AI systems in ways that are effective, efficient, ethical, and safe.
AI completes specific tasks based on your instructions
You and AI collaborate as creative thinking partners
AI works independently, guided by your vision & rules
Three modes of engagement — the 4Ds apply across all three.
Deciding what work to do yourself, what to do with AI, and how to divide it strategically.
Communicating clearly with AI — defining the what, how, and the style of interaction.
Evaluating AI outputs and behaviors with a critical, expert eye.
Ensuring AI interactions are responsible, transparent, and accountable.
These aren't tied to specific tools — they're durable competencies that grow with you.
Effective delegation requires understanding both what you're trying to accomplish and what AI can realistically do. The cornerstone of good delegation isn't about AI — it's about your own expertise.
Clearly define your goals and the nature of the work before involving AI. What does success look like?
Know the capabilities and limitations of the AI systems available to you — speed vs. depth, accuracy vs. creativity.
Strategically divide work — automate, augment, reserve for humans, or hand to agents — leveraging complementary strengths.
Description goes far beyond clever prompts. It's about building a thinking environment where both you and the AI can do your best work. AI can't read your mind — the quality of outputs depends on how clearly you articulate your needs.
Define what you want: the output, format, audience, style, and level of detail.
Guide how the AI approaches the task: methods, steps, data sources, analytical frameworks.
Set behavioral expectations: concise or detailed? Challenging or supportive? Explain reasoning or just deliver?
Discernment is the flip side of Description. If Description is about communicating what you want, Discernment is about deciding whether what you got back actually meets your needs. Even the most advanced AI benefits from human judgment.
Is this output accurate, appropriate, coherent, and does it truly solve the problem you intended?
Did the AI reason logically? Watch for attention lapses, circular reasoning, or reinserting rejected ideas.
Is the AI interacting with you effectively? Too verbose? Too brief? Responsive to feedback?
While the first three Ds address effectiveness and efficiency, Diligence addresses ethics and safety. It reminds us that AI interactions don't exist in a vacuum — our choices affect others.
Be intentional about which AI systems you use, what data you share, and how it aligns with values and policies.
Be honest and forthright about AI's role. Who needs to know? When should you disclose?
Take ownership of AI-assisted outputs. Verify facts, check for bias, and stand behind what you share.
The 4D Framework isn't meant to replace our existing approach — it's a lens to identify gaps and reinforce strengths. Consider the questions below as a team:
Which of the 4Ds map neatly onto competencies we already teach? Which represent genuine blind spots?
Would adopting terms like "Product Description" or "Process Discernment" give our team a sharper vocabulary?
What's one concrete change we could make next week that borrows from the 4D structure?
The 4Ds claim to be tool-agnostic and durable. Do we agree? Where might our framework need that same resilience?
Each table: share your single most valuable takeaway and one recommendation for how we might integrate a 4D concept into our existing framework.
One insight per group — what surprised you or challenged your assumptions?
Which "D" represents our biggest opportunity for improvement right now?
Name one specific, time-bound action step we'll take as a team in the next 30 days.
Document group responses — these become our working recommendations.
…emerge when humans and AI build on each other's strengths through Delegation, Description, Discernment, and Diligence.
Right work, right partner, right outcome.
Less wasted effort, faster iteration.
Transparent, fair, and values-aligned.
Private, secure, and accountable.
Based on the AI Fluency Framework by Anthropic, Ringling College, & University College Cork