By Jono Lowe, Founder of AI Literacy.org.uk
October 2025
During a recent panel discussion at the Sprite + Coventry University Thinktank event exploring AI in education, I found myself in polite disagreement with Dr Phil D Hall—a gentleman who helped invent chatbots and presumably knows a thing or two about them, certainly way more than me. His position against maintaining “the human in the loop” struck me as curiously pessimistic about teachers’ irreplaceable role in education. Whilst I deeply respect his technical expertise, my classroom experience suggests that educational chatbots work best precisely when teachers remain actively involved, guiding, monitoring, and intervening strategically.
This tension between technological capability and educational wisdom encapsulates the current moment in UK secondary education. Chatbots offer extraordinary potential for personalised learning support, yet their effectiveness depends entirely on thoughtful teacher implementation rather than autonomous operation. Schools embracing this technology with appropriate safeguards and teacher oversight discover transformative benefits for student learning, teacher efficiency, and educational equity. Frankly the pessimists and technophobes are going to be left behind.
What Educational Chatbots Actually Do
Educational chatbots represent a significant evolution from generic AI assistants like ChatGPT or Claude. Purpose-built platforms such as Flint AI (my personal favourite) create what’s termed a “walled garden”—safe, monitored environments where students interact with AI systems under teacher supervision and within carefully defined parameters.
Unlike open AI systems that students might use at home, educational chatbots operate within boundaries set by teachers. They can access only approved information sources (set ‘knowledge sources’ that the teacher sets up), provide responses aligned with curriculum requirements, and flag concerning interactions for teacher review. Most importantly, they create transparent records of student engagement that teachers can analyse for learning insights.
The technology sounds sophisticated, but implementation proves remarkably straightforward. I’ve recently trained department heads who describe themselves as “technophobic” to create functional educational chatbots within single training sessions, one did it under ten minutes. The barrier to entry is dramatically lower than many teachers anticipate—if you can write clear instructions and understand your curriculum, you can create effective chatbots.
These systems don’t replace teacher expertise; they extend it. A history teacher can create a chatbot that helps students analyse primary sources using specific analytical frameworks taught in class. An English teacher might develop one that guides essay planning through Socratic questioning rather than providing completed work. The chatbot becomes a tireless teaching assistant that operates according to the teacher’s pedagogical principles. Anecdotally, I have noticed that the ‘middle of the road’ quiet students use and benefit the most from chatbots as they are often the most reluctant to ask questions.
The Walled Garden Advantage: Safety and Oversight
One of the most compelling features of platforms like Flint AI is the comprehensive oversight they provide. Unlike students using unrestricted AI systems at home, educational chatbots create detailed interaction logs that teachers can review to understand student thinking, identify misconceptions, and evaluate learning progress.
This visibility transforms assessment possibilities. Teachers can observe how students approach problems, what questions they ask, where they struggle, and how they respond to guidance. It’s similar to having the ability to review every conversation during group work—insights that would be impossible to gather through traditional classroom observation alone.
The “walled garden” concept also addresses legitimate concerns about inappropriate content, privacy violations, and data security that accompany open AI use. Educational platforms implement age-appropriate content filtering, comply with GDPR requirements for UK schools, and provide institutional control over data management that generic AI services cannot match.
Safeguarding becomes proactive rather than reactive. Teachers can configure chatbots to flag concerning interactions—whether academic struggles suggesting intervention needs or language patterns indicating wellbeing concerns. Rather than discovering problems weeks later through traditional assessment, teachers receive early warning signals that enable timely support.
Perhaps most importantly for risk-averse school leadership, educational chatbot platforms provide audit trails and compliance documentation that satisfy both Ofsted requirements and parental expectations. Schools can demonstrate responsible AI use through systematic records of student interactions, teacher oversight, and learning outcomes.
Personalised Learning at Sustainable Scale
The holy grail of education—genuinely personalised learning for every student—has remained frustratingly elusive despite decades of effort and investment. Educational chatbots finally make this aspiration achievable within realistic resource constraints that UK secondary schools face. I would tentatively venture that Bloom’s “2-Sigma” problem is nearer to be resolved by embracing chatbots than ever before. Think about the implications of this.
Consider a typical Year 10 mathematics lesson with 30 students of varying abilities, prior knowledge, and learning speeds. Traditional teaching requires the teacher to pitch explanations at some mythical “average” student whilst trying to support struggling learners and extend high achievers simultaneously. It’s an impossible task that inevitably leaves some students under-served.
Educational chatbots transform this dynamic entirely. Every student can access personalised support calibrated to their specific needs, learning pace, and preferred explanation styles. A student who understands concepts quickly can accelerate through material whilst another who needs repeated explanations receives patient, varied approaches without holding back classmates or consuming disproportionate teacher time.
The chatbot doesn’t tire of explaining concepts repeatedly, doesn’t show frustration with students who struggle, and doesn’t inadvertently make students feel inadequate for needing additional support. For many students—particularly those who feel embarrassed asking questions in front of peers—this non-judgmental assistance proves transformative for learning confidence and academic progress.
Differentiation becomes systematic rather than aspirational. Teachers can configure chatbots with multiple explanation approaches, varied difficulty levels, and alternative examples that address different learning preferences. Rather than creating multiple versions of every resource manually, teachers design intelligent systems that adapt automatically to individual student needs.
Extending Teacher Impact Beyond Classroom Hours
One of the most practically valuable aspects of educational chatbots is their ability to provide consistent support outside traditional lesson times without requiring teachers to be constantly available. Students completing homework, revising for examinations, or working on extended projects can access subject-specific guidance whenever they need it.
This extended support particularly benefits students from disadvantaged backgrounds who may lack parental academic support, quiet study spaces, or resources for private tutoring. The chatbot provides expert assistance regardless of family circumstances, helping to level educational playing fields in ways that traditional interventions struggle to achieve.
Teachers describe the relief of knowing students have reliable homework support without needing to respond to emails at 10pm or field panicked messages before examinations. The chatbot handles routine questions about assignment requirements, concept clarification, and process guidance, freeing teachers to focus on more complex student needs that genuinely require human expertise.
Revision support becomes more structured and effective. Teachers can create chatbots that guide students through systematic review of course content, generate practice questions, explain marking criteria, and provide formative feedback on practice responses. Students can revise at their own pace, focus on areas where they struggle, and build confidence through repeated practice with immediate feedback.
Identifying Gaps and Misconceptions Systematically
Perhaps the most pedagogically valuable feature of educational chatbots is their ability to reveal student thinking in ways that traditional assessment rarely captures. Teachers can review conversation logs to understand how students approach problems, what assumptions they make, where misconceptions arise, and how effectively they respond to guidance.
This insight enables targeted intervention that addresses actual learning barriers rather than assumed difficulties. A teacher might discover that students struggle with a particular concept not because they lack foundational knowledge but because they misunderstand specific terminology or apply flawed reasoning in predictable ways.
The data becomes actionable at both individual and class levels. Teachers can identify students needing additional support before they fall seriously behind, recognise patterns of widespread misunderstanding that warrant whole-class reteaching, and evaluate whether their explanations effectively address common student confusions.
Assessment becomes more authentic and formative. Rather than relying solely on summative tests that reveal only final outcomes, teachers can observe learning processes, understand student reasoning, and provide guidance that addresses actual thinking patterns rather than just correcting wrong answers.
For students with learning differences or language barriers, chatbot interactions provide diagnostic information that helps teachers understand individual needs and calibrate support appropriately. The system creates detailed records of student capabilities, struggles, and progress that inform educational planning more comprehensively than traditional assessment alone.
Building AI Literacy Through Supervised Practice
Educational chatbots provide ideal environments for developing student AI literacy under expert guidance. Rather than experimenting with AI systems unsupervised at home—potentially developing problematic habits or unrealistic expectations—students learn to collaborate with AI effectively within structured educational contexts.
Teachers can model effective prompting strategies, demonstrate quality evaluation techniques, and explain when AI assistance is appropriate versus problematic. Students develop sophisticated understanding of AI capabilities and limitations through guided experience rather than trial and error.
This supervised introduction to AI collaboration prepares students for academic and professional futures where such skills will be essential. Universities and employers increasingly expect AI competence, but they want graduates who use these tools thoughtfully rather than dependently. Educational chatbots provide practice grounds for developing mature, critical approaches to AI assistance.
The metacognitive benefits prove substantial. Students who regularly interact with chatbots whilst reflecting on the experience develop better understanding of their own learning processes, clearer recognition of when they need help versus when they should persist independently, and stronger evaluation of information quality regardless of source.
Overcoming Teacher Workload Challenges
UK secondary teachers face chronic overwork that contributes to retention crises and burnout rates damaging both individuals and educational quality. Educational chatbots address several major workload pressure points whilst improving rather than compromising educational outcomes.
Routine question-answering consumes substantial teacher time both during lessons and through email exchanges. Chatbots handle these predictable queries efficiently, allowing teachers to focus attention on complex problems, pastoral support, and genuinely human interactions that AI cannot replicate effectively.
Homework support becomes scalable without requiring teachers to provide individualised assistance to 150+ students across multiple classes. The chatbot offers patient, detailed guidance to every student who needs it, whilst teachers focus on reviewing outcomes and addressing systematic issues rather than answering identical questions repeatedly.
Feedback provision improves in both quantity and quality. Teachers can configure chatbots to provide immediate formative feedback on practice work, freeing teacher time for more substantive commentary on summative assessments. Students receive more feedback overall whilst teachers spend less time on routine marking.
Resource differentiation becomes manageable rather than aspirational. Instead of manually creating multiple versions of every worksheet, lesson plan, and assessment, teachers design intelligent systems that adapt automatically to varied student needs. The initial investment pays dividends across repeated use and multiple cohorts.
The Human in the Loop: Why Teacher Oversight Matters
This brings me back to my disagreement with Dr Hall at the Coventry University event. His position against “the human in the loop” reflects, I suspect, confidence that AI systems will eventually operate reliably without human oversight. He certainly dazzled the attendees and I with his technical expertise. Whilst I admire that technological optimism, educational contexts demand more cautious approaches that prioritise student welfare over algorithmic efficiency.
Teachers provide essential judgement that AI systems cannot replicate: recognising when students need encouragement versus challenge, identifying pastoral concerns that manifest through academic struggles, understanding classroom dynamics that influence individual student behaviour, and making ethical decisions about appropriate support versus unhelpful dependency.
The “human in the loop” isn’t a temporary accommodation until AI improves—it’s a permanent feature of effective educational technology. Teachers curate content, set boundaries, interpret student interactions, and intervene when algorithms miss nuance or context that human expertise recognises immediately.
Moreover, students benefit from knowing teachers monitor their work. The oversight provides accountability that encourages effort, creates safety that enables risk-taking, and maintains relationships that sustain motivation through inevitable difficulties. Removing teachers from the loop would eliminate these vital psychological elements of effective education.
Educational chatbots work best not as autonomous systems but as tools that amplify teacher expertise and extend teacher capacity. The technology serves educational purposes defined by teachers rather than pursuing efficiency metrics disconnected from genuine learning outcomes.
Addressing Common Concerns and Misconceptions
Many teachers approach educational chatbots with understandable concerns about student dependency, academic integrity, and technological reliability. These worries, whilst legitimate, often reflect misunderstandings about how educational chatbots actually function within supervised contexts.
Student dependency concerns typically focus on chatbots doing work for students rather than supporting independent thinking. However, properly configured educational chatbots function more like Socratic tutors than answer providers—they ask questions, suggest approaches, and guide reasoning rather than supplying completed work.
Academic integrity questions arise from experiences with open AI systems that students might use to generate essays or complete assignments inappropriately. Educational chatbots address these concerns through teacher-defined parameters that prevent such misuse whilst encouraging appropriate collaboration that develops rather than bypasses learning.
Technological reliability worries often stem from early AI experiences where systems produced errors, inappropriate content, or unhelpful responses. Modern educational platforms incorporate substantial safeguards, continuous improvement based on teacher feedback, and oversight systems that catch problems before they affect students significantly.
Cost concerns prove largely unfounded. Many educational chatbot platforms offer free tiers or reasonable school licences that cost far less than traditional educational technology whilst providing greater functionality and flexibility. The return on investment becomes apparent quickly through improved outcomes and reduced teacher workload.
Implementation Success Factors
Schools achieving the greatest benefits from educational chatbots share several implementation characteristics that distinguish successful adoption from problematic deployments. Understanding these factors helps schools avoid common pitfalls whilst maximising technology benefits.
Teacher training proves essential but needn’t be extensive. As I’ve discovered training technophobic department heads, basic chatbot creation requires surprisingly little technical knowledge. Focus training on pedagogical design—what makes effective educational conversations—rather than technical operation.
Start small with pilot projects that prove value before attempting comprehensive implementation. Teachers who create single chatbots addressing specific pain points—homework support for one unit, revision guidance for upcoming examinations—build confidence whilst generating evidence that motivates broader adoption.
Collaborate across departments to share successful approaches and avoid duplicating effort. Subject-specific chatbots benefit from teacher expertise, but implementation strategies, technical troubleshooting, and best practices transfer readily across curriculum areas.
Establish clear policies about appropriate use, student expectations, and teacher responsibilities before deployment. Students need to understand when chatbot assistance is encouraged versus when independent work is required. Teachers need clarity about monitoring expectations and intervention protocols.
The Trajectory: Why Chatbot Effectiveness Will Only Improve
Educational chatbot technology remains relatively young, yet current capabilities already provide substantial benefits for teaching and learning. The trajectory suggests rapid improvement that will make these systems even more valuable for UK secondary schools in coming years.
AI capabilities continue expanding exponentially, with each generation demonstrating improved language understanding, reasoning capability, and response quality. Educational chatbots will inherit these improvements, becoming more helpful, accurate, and pedagogically sophisticated without requiring teacher retraining or system redesign.
Educational platforms evolve based on teacher feedback and classroom experience, incorporating features that address actual teaching challenges rather than pursuing technological novelty disconnected from educational needs. This user-driven development ensures platforms become more practically useful rather than more impressively complex.
Integration with existing school systems improves as educational technology providers recognise chatbot potential and develop compatible infrastructure. Future implementations will connect seamlessly with assessment systems, student information databases, and curriculum planning tools that schools already use.
Research evidence accumulates demonstrating chatbot effectiveness for specific learning outcomes, student populations, and subject areas. This evidence base will help schools make informed implementation decisions whilst providing Ofsted-acceptable documentation of educational impact and value for investment.
Subject and/or Topic-Specific Applications Across the Curriculum
Educational chatbots adapt to virtually any subject area, though specific applications vary based on disciplinary demands and pedagogical approaches. Understanding subject-specific possibilities helps teachers envision relevant uses within their own contexts.
In my own subjects (film and media studies), my students have greatly benefitted from using chatbots for each exam section, particularly from feedback on initial ideas. Young folk are surprisingly adept at responding positively to feedback. English and modern foreign languages benefit from conversation practice, writing support, and text analysis guidance. My hunch is that humanities subjects will benefit the most from this incredible tech. Chatbots can engage students in dialogue that develops language skills, provide feedback on draft writing without revealing complete solutions, and guide literary analysis through questioning and, as always with AI; iterate, iterate, iterate.
Mathematics and sciences use chatbots for problem-solving guidance, concept explanation, and procedural practice. The systems can work through examples step-by-step, identify where student reasoning diverges from correct approaches, and provide targeted hints that maintain challenge whilst preventing frustration.
Humanities subjects employ chatbots for research support, source analysis, and argument development. Students can explore historical evidence, geographical data, or philosophical positions through guided inquiry that develops critical thinking whilst ensuring accurate foundational understanding.
Creative subjects leverage chatbots for technique explanation, project planning, and reflective practice. The systems can guide students through design processes, suggest technical solutions to creative challenges, and facilitate the metacognitive awareness that distinguishes excellent from merely competent creative work.
Equity and Inclusion Benefits
Educational chatbots provide particular advantages for students who face barriers in traditional classroom environments. The technology addresses multiple equity challenges simultaneously whilst avoiding the stigmatisation that sometimes accompanies targeted support programmes.
Students with social anxiety or selective mutism can participate fully in learning conversations without the stress of public speaking or peer observation. The private nature of chatbot interaction removes social performance pressure whilst maintaining academic engagement and intellectual development.
English language learners benefit from patient language support that adapts to their current proficiency whilst maintaining age-appropriate intellectual challenge. Chatbots can provide translation assistance, clarify idioms or cultural references, and offer multiple phrasings that develop rather than bypass language development.
Students with learning differences such as dyslexia, ADHD, or processing disorders can access multi-modal support that accommodates their specific needs. Chatbots can read text aloud, break complex instructions into manageable steps, or provide visual representations that complement written explanation.
Disadvantaged students lacking home learning support receive expert assistance regardless of family circumstances. The chatbot provides the explanatory conversations, homework guidance, and revision support that advantaged students might receive from educated parents or private tutors.
Preparing Students for AI-Enhanced Futures
Perhaps the most important long-term benefit of educational chatbots is preparing students for academic and professional environments where AI collaboration will be standard expectation rather than novel innovation. Universities and employers increasingly assume AI competence whilst valuing thoughtful, critical approaches over dependent relationships.
Students who develop AI literacy through supervised school experience arrive at university better prepared for research, writing, and problem-solving that appropriately leverages AI assistance. They understand when AI helps versus hinders learning, how to evaluate AI outputs critically, and how to maintain academic integrity whilst using powerful tools effectively.
Professional preparation improves as students develop habits of effective human-AI collaboration that transfer directly to workplace contexts. Many careers will involve routine AI interaction, making supervised practice during secondary education valuable foundation for professional success.
Digital citizenship develops through guided reflection on AI capabilities, limitations, and ethical considerations. Students learn to think critically about algorithmic systems, recognise when human judgement should override automated recommendations, and consider societal implications of AI deployment across various contexts. Leveraging AI to develop critical thinking is, in my humble view, the key to maximising outcomes for students and if we start with the most disadvantaged students first, we can really help improve the future.
Practical Next Steps for Schools
Schools interested in exploring educational chatbot potential can begin with modest pilots that demonstrate value whilst building institutional confidence and capability. Success with initial projects typically generates enthusiasm that supports broader implementation.
Identify specific pain points where chatbot support might provide immediate relief—homework questions that consume disproportionate teacher time, revision support before high-stakes examinations, or concept explanation for topics where students consistently struggle. Target these challenges first.
Select teachers enthusiastic about experimentation rather than mandating universal participation initially. Early adopters generate evidence, develop expertise, and create internal champions who can support colleagues during broader rollout.
Choose platforms thoughtfully based on educational priorities rather than impressive features. Flint AI’s walled garden approach suits schools prioritising oversight and safety, whilst other platforms might offer advantages for specific subject areas or technical requirements.
Establish evaluation criteria before implementation to document benefits and identify areas needing improvement. Clear evidence of impact helps justify continued investment whilst demonstrating educational value to sceptical stakeholders.
My Personal Journey: From Sceptic to Advocate
I’ll admit—when I first encountered educational chatbots, I shared the instinctive technophobia that many teachers experience when confronted with new systems promising to “revolutionise education.” We’ve heard such claims before, usually accompanying expensive investments that gather dust after initial enthusiasm fades.
What changed my perspective was actually creating chatbots and observing student responses. The technology proved far more accessible than anticipated—I could design functional educational chatbots without programming expertise, technical support, or significant time investment beyond thoughtful pedagogical planning.
More importantly, students engaged with chatbots in ways they wouldn’t during traditional instruction. Reluctant participants asked questions freely, high achievers explored extensions independently, and struggling learners requested repeated explanations without embarrassment. The diversity of interaction patterns revealed learning insights I’d never captured through conventional assessment.
My perhaps excessive enthusiasm now—colleagues joke about my chatbot evangelism—stems from witnessing genuine benefits for students who previously struggled, teachers who found renewed professional satisfaction, and schools that discovered practical solutions to persistent challenges. The technology works, and it works remarkably well when implemented thoughtfully.
The Book: The Art of the Leverage
For non-teachers and teachers seeking comprehensive guidance on AI implementation beyond chatbots specifically, I’ve written The Art of the Leverage, exploring how we can all can harness AI tools effectively whilst maintaining human judgement and improving outcomes in a range of ways. The book combines practical implementation advice with philosophical reflection on AI-enhanced environments.
Self-published and available exclusively from Amazon, The Art of the Leverage provides step-by-step guidance on “prompt engineering” (and why it’s a misnomer), ethical considerations that support effective AI adoption. It’s written specifically for novices and ‘toe-dippers’ not just UK secondary teachers navigating these technologies whilst managing realistic time constraints and institutional expectations.
The book reflects my conviction that AI represents opportunity rather than threat for the committed- tools that can restore joy to teaching for example and improve outcomes in dozens of fields by eliminating drudgery whilst enhancing and augmenting rather than replacing the human relationships that make life meaningful.
Conclusion: Embracing Educational Chatbots Thoughtfully
Educational chatbots represent one of the most immediately practical AI applications for UK secondary schools—technology that delivers measurable benefits quickly whilst remaining manageable for teachers without technical expertise. The combination of accessibility, effectiveness, and safety makes chatbots ideal entry points for schools beginning AI integration journeys.
The benefits—personalised learning at scale, extended teacher impact, systematic identification of learning gaps, improved student outcomes, and reduced teacher workload—far outweigh the modest risks and challenges of thoughtful implementation. Schools investing in educational chatbots now position themselves advantageously for the AI-enhanced educational future that’s emerging rapidly.
My mild disagreement with Dr Hall at the Coventry University event ultimately reflects different philosophical commitments rather than technical understanding. Whilst I respect his expertise regarding chatbot capabilities, I maintain that educational contexts demand human oversight that prioritises student welfare over algorithmic efficiency. The human in the loop isn’t weakness—it’s strength.
Ready to explore how educational chatbots can transform learning in your school? I’m available to deliver bespoke AI in education training across the UK, including hands-on workshops where even technophobic department heads learn to create effective educational chatbots. From initial pilots to whole-school implementation, we can design professional development that delivers immediate results. Email jono@ai-literacy.org.uk to discuss how chatbot technology can address your specific teaching challenges whilst enhancing student outcomes.
For systematic, accredited professional development, the NCFE Level 2 Certificate in AI Literacy provides comprehensive training designed specifically for UK educators. This regulated qualification ensures your staff develop the knowledge and confidence needed to implement AI effectively whilst maintaining educational excellence and professional standards.