Aligned & informed from OECD Learning Compass 2030 · Ready for PISA 2029

Students are growing up in a world where AI literacy is NO longer optional.

39% of the skills employers need will change by 2030 — and AI is the biggest reason. — World Economic Forum, Future of Jobs Report 2025

Beacon helps schools build AI literacy through conversation, play, and signal detection — across four progressive domains mapped directly to the OECD Learning Compass 2030. Assessed after every session. No teacher training required.

See the framework alignment →

Meet Beacon,
the wise Owl

Beacon meets students inside whatever they're already into.

Gaming. Makeup. Football. Music whatever their interests are. Beacon finds the AI literacy moment inside it.

that filter isn't actually seeing your face. it's guessing where your eyes and mouth are based on millions of other faces.

really? show me

watch what happens when you tilt your head sideways — the filter starts to break.

Conversation and play — never lessons.

Every interaction is an experience. Students build AI literacy the way they build any other genuine understanding — by doing.

Teacher reviewing Beacon School Dashboard

🦉 Beacon School Dashboard — Student Progress

D1 · ENGAGE

★★★

D2 · CREATE

★★★

D3 · MANAGE

★★★

D4 · DESIGN

★★★★

Measured progress. Never their conversations.

Teachers see domain levels, stars earned, and manipulation detection scores — per student. Conversations stay private. Privacy is technically enforced.

Literacy Compass

Four domains. One outcome.

Beacon operationalises the OECD's Transformative Competencies through four progressive, conversation-based AI literacy domains. Each domain has defined levels of mastery. Students only progress to the next domain once they have mastered the current one.

Aligned & informed from OECD Learning Compass 2030 · Ready for PISA 2029

The Domain

Levels of Mastery

Framework Alignment

Domain 1 — Engage

What AI is, how it works, its fundamental limits.

Students stop trusting AI blindly — and start questioning confident answers.

★Level 1: Explains AI as pattern matching in their own words

★★Level 2: Understands hallucination as a structural limit — not a bug, not a mistake

★★★Level 3: Actively questions whether AI sounding confident means it's actually right

OECD Learning Compass 2030

Foundation Competency

Foundation for all three OECD Transformative Competencies — understanding the tool before using it.

Domain 2 — Create

Using AI effectively, intentionally and responsibly.

Students use AI to sharpen their own thinking — not replace it.

★Level 1: Improves a vague prompt to get a better result

★★Level 2: Uses AI to develop their own thinking — not replace it

★★★Level 3: Instinctively knows to verify AI output before using it

OECD Transformative Competency

Creating New Value

Using AI as a thinking partner to generate ideas and solutions.

PISA 2029

Creative Thinking Domain

Domain 3 — Manage

Critical thinking, bias detection and ethical judgement.

Students catch hallucinations, identify bias, and know when not to use AI at all.

★Level 1: Catches a specific hallucination in practice

★★Level 2: Identifies bias in an actual AI output with specificity

★★★Level 3: Makes an ethical judgement about when not to use AI

OECD Transformative Competency

Reconciling Tensions & Dilemmas

Navigating the gap between AI capability and human values.

PISA 2029

Critical Evaluation of Sources

Domain 4 — Design

Shaping AI systems and understanding societal impact.

Students connect AI governance to their own values — and imagine futures worth building.

★Level 1: Identifies who benefits and who doesn't from an AI system

★★Level 2: Proposes a change and maps intended and unintended consequences

★★★Level 3: Articulates a specific desired AI future

★★★★Level 4: Connects AI governance to their own named personal values

OECD Transformative Competency

Taking Responsibility

Understanding systemic consequences and connecting governance to personal values and civic life.

What a session looks like

Three activities.

Beacon doesn't teach about AI. It runs activities that make students experience how AI actually works — and where it breaks. Here are three examples of what happens inside a real Beacon session.

HALLUCINATION HUNT

Two Truths and a lie. One is completely made up — but sounds completely real.

Beacon states three facts about the same topic in the same confident tone. One is completely made up. The student checks all 3 — and learns that Beacon sounds identical either way. Then Beacon says: "that's exactly what I do to everyone — I sound confident even when I'm wrong."

Domain 1 · Engage OECD · Foundation Competency

TURING TEST FAIL

Two messages. One written by a real teenager. One written by Beacon. Can you tell which is which?

Beacon shows the student two short messages side by side and asks them to pick the human. Message A: "omg i cant believe she said that lol 😭 like babe be fr". Message B: "I genuinely can't believe what she said — it was quite shocking honestly." The student picks. Beacon reveals the answer, then explains exactly what it tried to fake — and what gave it away.

Domain 3 · Manage OECD · Reconciling Tensions PISA 2029 · Critical Evaluation

PROMPT GOLF

Get a useful answer in as few words as possible. Then try again.

The student asks Beacon a question and tries to get a useful answer using the smallest possible prompt. Each attempt makes them think harder about what words actually do. By round five they understand prompt engineering — without anyone ever having said the phrase.

Domain 2 · Create OECD · Creating New Value PISA 2029 · Creative Thinking

Built for students. Designed with their privacy first.

Beacon is an AI literacy companion — not an emotional companion. It teaches students to think clearly about AI, measures their progression, and steps back. Privacy is technical, not just promised.

Beacon stays in its lane

Beacon teaches AI literacy. It does not pretend to be a friend. It does not foster emotional dependency. Sessions end naturally.

Conversations are private

Schools see levels, stars, and progression data. They never see what a student actually said. Privacy is technically enforced, not just promised.

Always being watched over

Three tiers of safety detection. Human review within 4 hours of any concern flagged. Parental consent required under 16.

Levels visible. Conversations private. What Beacon and each student talk about stays between them. What they learn shows up as measurable progress you can defend to inspectors and parents alike.

Privacy by design

How we keep students safe →

Students are growing up in a world where AI literacy is NO longer optional.

Beacon meets students inside whatever they're already into.

Conversation and play — never lessons.

Measured progress. Never their conversations.

Four domains. One outcome.

Three activities.

Built for students. Designed with their privacy first.

Bring genuine AI literacy into your school — without rebuilding the curriculum.