AI Critical Thinking Education: Addressing Bias in Classroom AI

Almost 30% of Saudi teachers already correct biased AI outputs, highlighting an urgent need for students to interrogate AI, not just trust it.

In this episode:

• A 2025 Saudi Arabia survey found nearly 30% of teachers are already correcting AI bias in education, highlighting an urgent need for students to interrogate AI.
• The core issue: most AI tools are trained on English-language and Western-dominant datasets, creating linguistic and cultural blind spots in AI-generated knowledge for diverse learners.
• Teachers must evolve into 'epistemic intermediaries,' guiding students in AI critical thinking education by modeling how to assess AI outputs for accuracy and cultural relevance.
• True AI literacy for students involves collaborative reasoning and actively critiquing AI responses, not just passively accepting them.
• Designing assessment tasks around "Product, Process, and Performance" can ensure students engage in cognitive stretch, applying unique context and judgment, which cannot be faked by AI tools.

Chapters:

• 00:00 — Cold open & welcome
• 00:25 — Saudi Arabia's AI bias challenge: 30% of teachers correcting AI
• 00:55 — Cultural and linguistic blind spots in AI tools
• 01:50 — AI critical thinking education: shifting from teacher as authority to AI interrogator
• 02:45 — The teacher's new role: 'epistemic intermediary' assessing AI outputs
• 03:50 — Redefining AI literacy for students: collaborative reasoning and critique
• 04:45 — Saudi Arabia's proactive approach to addressing AI bias in education
• 05:25 — School leaders: prioritizing AI critical thinking education over technology adoption
• 06:10 — Protecting human judgment, imagination, and wisdom in responsible AI in education
• 06:45 — Knowledge transmission to knowledge interrogation: The core shift

How can teachers address AI bias in education in their classrooms?
Teachers can address AI bias by becoming 'epistemic intermediaries,' systematically assessing AI-generated content with students for factual accuracy, linguistic precision, cultural relevance, and contextual appropriateness.

What does AI critical thinking education look like for students?
AI critical thinking education involves teaching students to systematically critique AI responses, compare outputs across languages, identify inconsistencies, and consciously inject missing cultural nuance into AI-generated content.

Why is responsible AI in education crucial for non-English dominant contexts?
Responsible AI in education is crucial because most AI tools are trained on English-language and Western-dominant datasets, leading to inherent linguistic and cultural blind spots that can misrepresent local realities for students in other regions.

Featuring: Dan Fitzpatrick, Basmah AlBuhairan, Reem Taibah, Amani AlOlayani, King Abdulaziz City for Science and Technology, Centre for the Fourth Industrial Revolution Saudi Arabia, Ministry of Education, Saudi Arabia, World Economic Forum.

Follow AI in Education with Dan Fitzpatrick for more on AI in education.

Transcript

SPEAKER_00 0:00

If this episode makes you think, please let us know in the comments and support us by subscribing and leaving a review. Thank you. Today we are exploring a fascinating piece from the World Economic Forum published on June 22nd, 2026, titled As AI in the Classroom Becomes a Mainstay, Teaching Critical Thinking Becomes Essential. And here's the hook. A 2025 survey across Saudi Arabia involving nearly 45,000 teachers found that almost 30% were already actively correcting or adapting biased AI outputs in their teaching. That's a significant number, isn't it? It really highlights the urgent reality educators are facing right now. The article itself is incredibly insightful, co-authored by Basma Al-Buhiran, who's a senior advisor to the president of King Abdullah's City for Science and Technology, along with Reem Taiba and Amani Alullayani, both project leads from the Center for the Fourth Industrial Revolution, Saudi Arabia. They're looking at how artificial intelligence is changing the classroom, especially through the lens of cultural and linguistic context, and what that means for teaching students to think critically. Now the core argument they lay out is pretty stark. Most large language models, those amazing AI tools we're all playing with, are predominantly trained on English language and Western dominant data sets. What this means practically is that they have these inherent linguistic and cultural blind spots. When students use these tools to summarize information, generate ideas or answer questions, the AI generated knowledge often doesn't fully capture the richness of other languages or the depth of local cultural context. Think about a year eight geography lesson in, say, a school in Cairo or Kuala Lumpur. An AI might generate examples or frame concepts in a way that just doesn't resonate with the local realities or cultural nuances for those students. The piece makes the case that this isn't just about bad translations, it's about whether the AI generated information genuinely reflects the world of the learners engaging with it. This brings us to the first big shift for us as educators. For so long, the teacher has often been the primary unquestioned authority in the classroom. We're the ones who deliver, interpret, and validate information. Students are conditioned to trust that epistemic authority. But what the authors, Albu Hiran, Tiber, and Alulayani, are worried about, and what I really found myself nodding along to is the growing risk that this dynamic simply transfers to AI. Students might start treating AI responses as authoritative answers, rather than outputs shaped by training data that, as the article points out, can carry linguistic gaps, cultural bias, and Western-centric assumptions. This is where AI critical thinking education becomes absolutely essential. We're not just giving them tools, we're teaching them how to interrogate those tools. The second thing this report highlights is how this changes our role as teachers. Instead of being the sole purveyors of knowledge, we become these vital epistemic intermediaries, as the article puts it. This means systematically assessing AI-generated content right there in front of our students, checking for factual accuracy, linguistic precision, cultural relevance, and contextual appropriateness. Imagine a Year 10 history class where students are asked to use AI to research a local historical event. The AI might provide a general overview, but it might completely miss the specific local context or historical interpretations that are vital. Our job then isn't just to point out the inaccuracies, but to model how we identify those gaps, how we question the source, and how we dig deeper to find the truth that AI might have missed. It really moves the teacher from content delivery to content curation and critical guidance. It's about helping students understand that the real value is not in what the machine produces, but in how the student responds to and engages with that output. For those of you listening, if you're finding these conversations valuable for your school or classroom, please do hit that follow button. We're always trying to bring you the most compelling and insightful perspectives on AI in education, so you don't miss out on these important discussions. Now, let's talk about what this means for AI literacy for students. It's not about students learning to code or memorizing the latest AI features. As I've always argued, it's about collaborative reasoning ability. It's about thinking with AI, not just using it as a black box. The article suggests encouraging students to systematically critique AI responses, to compare outputs across languages, and to identify inconsistencies. This is a game changer. For a year seven English class writing a short story, they might use AI to generate character descriptions. Instead of just accepting them, the teacher could ask them to evaluate how culturally appropriate those descriptions are or how they might be biased towards certain stereotypes, then challenge them to rewrite the AI's output, consciously injecting the cultural nuance that was missing. This moves learning from passive consumption to active productive struggle. It truly requires them to outsource their doing, not their thinking. The authors also touch on the exciting work happening in Saudi Arabia, which they suggest offers an early signal for the world. Especially for non-Englominant education systems. The Ministry of Education in Saudi Arabia, working with the Center for the Fourth Industrial Revolution, Saudi Arabia, and King Abdulaz's city for science and technology is actively fostering this awareness of linguistic and cultural bias in AI. That survey of 44 Thado's 920 teachers is concrete proof that educators are already on the front lines, adapting and evolving their practice. This isn't some distant future scenario, it's happening right now. They're demonstrating that teachers, when given the space and support, can become the best drivers of innovation. It anchors AI to existing friction points, like finding culturally relevant resources and empowers teachers as true change agents. For school leaders, this really underscores the purpose over technology principle. It's not about rushing to get the latest AI tools into every classroom, but about defining the learning goal first. The learning goal here is clearly AI critical thinking education. How do we design learning that cannot be faked because it demands depth, care, and imagination? It means designing assessment tasks that require cognitive stretch, where students have to apply their unique context, perspective or judgment, rather than just recall information that an AI could easily generate. We might ask students not just for a product, but also for the process. Show me your AI interaction logs. How did you prompt the AI? What biases did you identify? How did you correct them to make the output culturally relevant? This brings in the three Ps of assessment product, process and performance. The ability to question information, to identify AI bias in education, and to guide students through that evaluation process is rapidly becoming one of the most essential skills in education. It protects those irreplaceable human domains, judgment, imagination, and wisdom. Machines can compute, but they cannot wonder, and they cannot care about the nuances of a student's cultural heritage. This evolution, not revolution, of the teacher's role is exciting. It allows us to hold the complexity, so we have capacity for creativity and to ensure that responsible AI in education is not just a buzzword, but a lived reality in every classroom. The real shift here is from knowledge transmission to knowledge interrogation. That's all for today. Thanks for listening.