Cambridge Just Added 'Oracy' and AI Literacy to Its Framework — What That Means for ESL Class
On 7 July 2026, Cambridge University Press & Assessment updated its Life Competencies Framework — the model it's used since 2018 to describe transferable skills alongside language learning — to add four new focus areas: AI literacy, wellbeing, oracy, and global citizenship. "Effective language teaching should mirror the world students are living in," said Francesca Woodward, Cambridge's Global Managing Director for English, in the announcement. That's a fairly ordinary-sounding line for what's actually a notable admission: a major exam board is now saying that knowing grammar and vocabulary isn't the finish line — being able to speak spontaneously and evaluate what you're told, including what an AI told you, is now part of the definition of a competent English user.
Two of those four new areas — oracy and AI literacy — map onto something ENK has been building since before this framework update existed. Not because anyone predicted Cambridge's move, but because unscripted spoken output and structured evaluation of claims were already the whole point of a chunk of ENK's app library. That's the practical center of this piece: AI literacy activities for ESL classrooms don't have to wait for a dedicated AI curriculum to show up, because a handful of ENK's existing speaking and evaluation apps already train the transferable half of the skill. Here's what actually changed in the framework, and which of ENK's apps genuinely train each half of it.
What actually changed in the framework
The Cambridge Life Competencies Framework (CLCF) has always organized itself around six core competencies — Creative Thinking, Critical Thinking, Learning to Learn, Communication, Collaboration, and Social Responsibilities — resting on three foundation layers: emotional development, digital literacy, and subject knowledge. The July update doesn't replace that structure; it widens what counts as "communication" and "digital literacy" for a post-pandemic, AI-accelerated context.
The AI literacy piece is graded by age, which is worth reading closely because it's more specific than the usual "teach kids about AI" advice. Primary-age learners are expected to identify fake AI-generated images. Secondary students move to spotting bias in AI-generated texts. Adult learners are asked to recognize inaccuracies in workplace AI outputs. That's a genuine progression in critical evaluation, not a single skill repeated at three difficulty settings. Cambridge frames the stakes plainly: "Being able to identify and discuss AI content is a critical life skill." The update also names something less often said out loud in ELT materials — that both teachers and learners are struggling with the pace of AI's arrival, and the framework treats that anxiety as a wellbeing issue worth naming, not just a skills gap to close.
Why "oracy" isn't a new word for ESL teachers, even if it's new to this framework
Oracy — structured spoken competence, treated as seriously as literacy — isn't a term Cambridge invented, but folding it into a life-competencies framework signals something: spoken fluency assessed only through exam-style question-and-answer isn't the same skill as sustaining spontaneous, unscripted speech in front of other people. That distinction is the whole design premise behind two of ENK's speaking apps.
Act it Out (A1–C1) generates a role-play scenario with a named role for every performer, then can inject a plot twist mid-scene that performers have to absorb in real time, using language they have available rather than language they've rehearsed. Littlewood (1981) argues role-play lowers L2 anxiety specifically because it hands students a protective persona — they can attribute the language to their character, not to themselves — which makes risk-taking, and therefore genuinely spontaneous speaking practice, more likely than a direct question-and-answer format would.
Iconic Stories (B1–C1) gives a small group a CEFR-calibrated starter sentence built around a set of icons and a target grammar structure, then has students take turns extending the narrative live, no notes allowed, until every icon has appeared in the story. Swain's (1985) Output Hypothesis is the theory behind why this works: sustained, unrehearsed production forces learners to move from recognizing language to actively generating it, which is a different cognitive load than filling in a worksheet blank. Both apps put students in a position where the next sentence has to come from them, on the spot — which is a fair working definition of oracy, whether or not either app was built with that word in mind.
Teaching students to evaluate AI-generated content is a critical thinking exercise ENK already runs — just not badged as "AI"
The harder half of Cambridge's update is the evaluative piece: can a student look at a claim, a text, or an image and reason about whether it's accurate, biased, or worth trusting? That's not really an AI-specific skill — it draws on the same general critical thinking skills ESL classrooms already train, ones that AI content happens to make more urgent. Say Mean Quotes (B2–C2) was built around exactly that muscle, using the "Say, Mean, Matter" analytical framework from critical literacy research. Fisher & Frey (2012) identify text-based evidence, inference, and connection as the three core moves of critical text analysis, and the app sequences students through all three before they're allowed to react personally. One built-in activity, "Author Blind Spot," hides the quote's source until after the interpretation stage, then reveals it and asks directly: does knowing who said this change how you read it? That's a confirmation-bias check, run on a real quote instead of an AI-generated one, but it's the same reasoning move Cambridge's secondary-level "spot bias in AI texts" skill is asking for.
Reaction Reactor (A2+–C1) works the same muscle from a different angle — a surprising or provocative statement drops, and students have to produce a genuine, tonally appropriate reaction: shock, doubt, sarcasm, agreement, rather than defaulting to a flat "I think..." every time. Judging how much doubt a claim deserves, and having the actual language ready to express that doubt, is the pragmatic half of evaluating AI-generated content — knowing something's off is only useful if a student can say so in English, in the moment, at the right register.
A simple sequence to try this week
None of these four apps needs a special "AI literacy unit" wrapper — they slot into a normal lesson in order of rising cognitive demand:
- Say Mean Quotes — practice the Say-Mean-Matter sequence on a real quote, ending with the Author Blind Spot reveal to make source evaluation explicit.
- Reaction Reactor — drill the language of doubt, surprise, and disagreement so students have something to say once they've decided a claim looks shaky.
- Act it Out — put students in a scenario, then twist it mid-performance, so they have to think and speak on their feet rather than recite.
- Iconic Stories — close with a fully collaborative, notes-free narrative that forces sustained spontaneous production from everyone in the group, not just the confident talkers.
I ran the first two of these back to back with a B2 adult class the week the Cambridge update came out, using a genuinely ambiguous quote for the Say Mean Quotes round. The Author Blind Spot reveal landed harder than I expected — two students who'd argued confidently for the quote's message during the Mean stage visibly recalculated once they saw who'd actually said it, and one asked, unprompted, "wait, does that mean it's less true, or just less trustworthy?" That's precisely the distinction Cambridge's AI-literacy skill is asking secondary and adult learners to hold onto, and it surfaced from a discussion prompt with no AI content in it at all.
What this doesn't replace
To be direct about the limits here: none of ENK's apps generate or display actual AI images for students to fact-check, so the primary-level "spot the fake AI image" skill Cambridge names isn't something these four apps train directly — that still needs dedicated media-literacy material. What ENK's apps do build, reliably, is the underlying spontaneous-speaking practice and the habit of pausing to ask "should I trust this?" before reacting — the transferable half of the skill, not the AI-specific content layer sitting on top of it. Pair the two rather than expecting one to stand in for the other.
All four apps sit inside the same account as ENK's other 22 apps, with 100 AI credits a month at no cost, or from €5/month for 300 credits on the pricing page — no separate tier for these formats. Browse the full app library to see how they sit alongside ENK's other Conversation & Creativity apps. For more on the spoken-interaction side of this, see peer-to-peer speaking activities; for the argument-evaluation side, see how to run a classroom debate in English.
FAQ
What is "oracy" and why is Cambridge adding it to its framework now? Oracy means structured spoken competence — treated with the same seriousness as reading and writing, not just informal chat. Cambridge added it in its July 2026 Life Competencies Framework update to reflect that spontaneous spoken communication is a distinct, teachable skill separate from exam-style question-and-answer fluency.
What does "AI literacy" mean for ESL students specifically? Per Cambridge's update, it's an age-graded skill: primary students learn to spot fake AI-generated images, secondary students learn to identify bias in AI-generated texts, and adult learners practice recognizing inaccuracies in workplace AI outputs. It's a critical-evaluation skill, not a technical one.
Which ENK apps train spontaneous speaking practice specifically? Act it Out (A1–C1) and Iconic Stories (B1–C1) are the two built around unscripted, live spoken output — the first through improvised role-play with a mid-scene twist, the second through collaborative storytelling with no notes allowed.
Do any ENK apps teach students to evaluate AI-generated content directly? Not the AI-specific content layer — ENK doesn't display AI images or texts for students to fact-check. But Say Mean Quotes and Reaction Reactor train the transferable reasoning and language of evaluating and reacting to a questionable claim, which is the same muscle Cambridge's AI-literacy skill depends on.
Written by Prof. Vito Schiuma, designer of English, no kidding.