Practical AI education for people and organizations.

What we teach

We teach durable skills: how to frame a task, choose the right tool, verify results, protect sensitive information, and keep human judgment in charge.

A method that adapts to the learner

AI43 courses are guided by an adaptive multi-agent learning system. It combines the authors’ decade of teaching experience with active practice, frequent feedback, reflection, and progressive challenge — and responds to each learner’s pace. People remain responsible for the curriculum and its quality.

Evidence-informed learning design

The technology rests on a clear instructional foundation. We use established models as practical design tools—not as labels—and adapt them to the subject, audience, and intended outcome of each course.

Backward Design — Grant Wiggins & Jay McTighe

We define the intended outcome first, decide what observable work will demonstrate it, and only then select the content, practice, and feedback needed to get there.

First Principles of Instruction — M. David Merrill

Learning is organised around a meaningful real-world problem. Each path activates prior knowledge, demonstrates a new approach, gives the learner a chance to apply it with feedback, and helps integrate the result into their own work.

Universal Design for Learning — CAST

When developing materials, we take account of CAST UDL Guidelines 3.0: reducing unnecessary barriers and offering purposeful options for engagement, representation, action, and expression.

Three agents, one learning experience

Each agent has a distinct role, so the experience feels personal while responsibility stays clear.

Coach Agent

Your personal mentor. It explains the purpose of each quest, assigns the task, reviews your solution carefully, and gives focused feedback—not a ready-made answer, but the next useful step.

Teammate Agent

Your virtual classmate. It asks useful and sometimes uncomfortable questions, offers another approach, shares its own solution, and creates the momentum of a live group even when people study at different times.

Methodist Agent

The invisible course architect. It reads progress through the CIDPP framework, spots gaps, adapts the sequence, and helps keep material current as new AI models appear. It supports the authors; methodological decisions and responsibility remain with the AI43 team.

CIDPP: a framework for learning quality

Every learning unit is reviewed through five complementary criteria. Together they keep the course clear without becoming shallow, and rigorous without losing contact with real work.

C — Clarity

Clear goals, explanations, expectations, and structure.

I — Integrity

A coherent and sufficiently complete treatment of the material.

D — Depth

Enough depth to develop judgment, not only repeat a recipe.

P — Practicality

Direct usefulness in realistic work and everyday situations.

P — Pertinence

A fit with the learner’s level, needs, goals, and context.

Technology: Teleteacher

Teleteacher is the platform AI43 uses to run courses directly inside Telegram. Learners receive quests, respond, compare approaches, work with the agents, and move through an adaptive path. It is not a feed of theory — it is structured immersion, where AI is both the subject of study and part of the learning environment.

Access is part of the mission

We keep tuition affordable and offer subsidized and free programs for individuals, nonprofits, and media organizations facing barriers to learning.

Who we are

AI43 is a distributed team of educators, practitioners, and AI agents working from Amsterdam, Barcelona, Rome, and Dubai, with data centres in Canada and the Netherlands. People set priorities, make decisions, drink coffee, and take responsibility; AI agents support research, learning materials, and operations. Time zones are not a bug—they are part of the system.