Three Governments, One Blind Spot

In the span of one week, three separate institutions released frameworks for AI literacy. The US Department of Labor. The UK government. And a massive Brookings Institution study spanning 50 countries. All three are serious, well-researched, and well-intentioned. And all three have the same hole.

They teach people to understand AI. They teach people to use AI. They teach people to evaluate AI output. What none of them teach is how to communicate through AI systems — the skill that actually determines who thrives and who gets left behind.

Framework #1: The DOL's AI Literacy Framework

On February 13, the US Department of Labor released its AI Literacy Framework with five content areas:

1. Understanding AI principles
2. Exploring AI uses
3. Directing AI effectively
4. Evaluating AI outputs
5. Using AI responsibly

It's a clean taxonomy. Each area maps to real competencies. “Directing AI effectively” even nods toward prompt engineering. But notice what's assumed: the user sits outside the system, issuing commands and evaluating results. The framework treats AI as a tool you operate, not a medium you communicate through.

This is like teaching someone to read without teaching them how a library works — or more precisely, like teaching someone vocabulary without teaching them conversation.

Framework #2: Brookings' “Prosper, Prepare, Protect”

The Brookings Institution's global study is staggering in scope: 500+ stakeholders, 50 countries, 400+ studies reviewed. Their finding is blunt — AI risks to students “currently overshadow benefits” and “undermine foundational development.”

The three pillars — prosper, prepare, protect — frame students as recipients of AI's effects. Prosper means capturing upside. Prepare means building skills. Protect means mitigating harm. All three position the student outside the system, as if AI is weather: something that happens to you, not something you participate in.

But the risk Brookings identifies — overreliance on AI tools undermining development — isn't a tool problem. It's a communication problem. Students who can't navigate the application layer don't overrely on AI because they're lazy. They overrely because they lack the communicative fluency to engage with these systems critically. They can't tell when the system is giving them what they asked for versus what they need.

Framework #3: The UK Computing Curriculum Overhaul

On February 10, the UK government announced a major overhaul of its computing curriculum, concluding the current framework is “too narrow.” The new curriculum will emphasize digital literacy, AI awareness, and critical thinking alongside traditional programming.

This is genuine progress. The UK is one of the first countries to officially recognize that computing education can't just be coding anymore. But the additions — AI awareness, digital literacy, critical thinking — are still framed as knowledge about systems. Know what AI is. Know what it can do. Think critically about its outputs.

What's missing is the communicative layer: how do you actually talk to these systems? How do you navigate the application layer — the APIs, the interfaces, the prompt structures, the feedback loops — as a communicative environment rather than a set of buttons to press?

The Pattern

Three frameworks. Three different institutions. Three different scopes (workforce, global education, national curriculum). All converging on the same model:

This model treats AI systems as objects. You learn about them. You learn to operate them. You learn to judge their output. At no point do you learn to inhabit the communication space they create.

This is the gap that Application Layer Communication (ALC) was built to name.

What ALC Adds

ALC reframes the relationship between humans and software systems as fundamentally communicative. You're not just “using” an AI tool. You're communicating through the application layer — the space where interfaces, APIs, prompts, and algorithmic responses create a conversational environment.

Under ALC, the DOL's five areas don't disappear — they become prerequisites. Understanding AI principles is necessary background. Directing AI effectively is a communicative skill, not just a technical one. But literacy without fluency is incomplete. You can know every word in a language and still fail at conversation.

The critical addition is ALC fluency: the ability to navigate the application layer as a communicative environment. This includes:

• Recognizing communicative registers — different systems require different “tones” (prompting ChatGPT ≠ querying a database ≠ configuring an API)
• Reading system feedback as dialogue — error messages, partial completions, and unexpected outputs are responses in a conversation, not failures
• Navigating between layers — moving from a GUI to an API to a configuration file is code-switching, not escalation
• Building folk theories as proto-communication — your mental model of how the algorithm works isn't failed knowledge, it's the beginning of a communicative relationship

Why This Matters: The Stratification Problem

The practical consequence of the blind spot is stratification. When frameworks teach AI literacy without ALC fluency, they create two tiers:

• High-ALC users who can navigate system boundaries, switch registers, and extract value from AI systems across contexts
• Low-ALC users who know about AI but can't communicate through it effectively — stuck in whatever interface they were trained on

The DOL framework, for example, will produce workers who can use AI tools in their specific workflow. But when the tool changes — when the API updates, when the interface redesigns, when a new system replaces the old one — only those with ALC fluency will transfer their skills. Everyone else starts over.

This is how digital inequality reproduces itself. Not through lack of access (though that matters), but through lack of communicative fluency in the spaces where value is created.

What Would an ALC-Informed Framework Look Like?

An AI literacy framework that includes ALC would add at least two dimensions to the standard model:

The first addition — navigating the application layer communicatively — means understanding that you're not just pressing buttons. You're in a dialogue. The system responds to how you communicate, not just what you ask.

The second addition — transferable fluency — is what separates literacy from fluency. A literate user can operate one tool. A fluent communicator can move between systems, adapting their approach the way a multilingual person switches languages.

The Convergence Is the Opportunity

The fact that three major institutions independently arrived at the same model — and the same blind spot — in the same week isn't failure. It's an opportunity. The policy world is hungry for AI literacy frameworks. They're building them fast. The communicative dimension is missing not because anyone rejected it, but because no one has named it yet.

ALC names it. And the gap between “AI literacy” and “ALC fluency” is exactly where the next wave of digital inequality will form — or where it gets prevented.

Sources:
US Department of Labor AI Literacy Framework (Feb 13, 2026) — Axios
Brookings Institution, “A New Direction for Students in an AI World: Prosper, Prepare, Protect” (Feb 2026) — Brookings
UK Computing Curriculum Overhaul (Feb 10, 2026) — gov.uk