The Resistance Routine: Engineering 'Desirable Difficulty' with AI to Secure Grade 9 and A* Results

The Fluency Trap: Why Your Revision Feels Productive But Isn't
It is a common scene during the Easter holidays: a GCSE or A-Level student surrounded by pastel highlighters, beautifully condensed mind maps, and a series of AI-generated summaries. Everything looks perfect. The student feels confident. They have 'covered' the specification. But when they sit down for a mock paper, the knowledge seems to evaporate. The 'Grade 9' clarity they felt while reading vanishes the moment they are faced with a blank page.
This is what educational psychologists call the 'Fluency Trap' or the 'Illusion of Competence'. When information is easy to process—when it is presented in a clean, AI-generated summary or a pre-made Flashcard deck—your brain assumes it has mastered the material. In reality, you have only mastered the art of recognising the material. To move knowledge from short-term recognition to long-term retrieval, you need 'Strategic Friction'.
What is Strategic Friction?
Robert and Elizabeth Bjork, leading researchers in cognitive psychology, pioneered the concept of 'Desirable Difficulties'. Their research proves that the more effortful the learning process, the more durable the memory. If revision feels easy, you are likely wasting your time. If it feels frustrating, slow, and mentally taxing, you are likely learning.
In the age of generative AI, most students use technology to remove friction. They ask AI to 'explain this simply' or 'summarise this chapter'. While this helps with initial understanding, it is the enemy of high-level exam performance. To secure the top grades in 2025 and 2026, you must flip the script. You must use AI not as a shortcut, but as a 'Difficulty Engine'.
Phase 1: The 'Devil's Advocate' Protocol (A-Level Humanities)
For subjects like History, RS, or English Literature, the jump from a Grade B to an A* often depends on the strength of your evaluation. Most students use AI to find supporting evidence for their arguments. This makes the revision process too 'smooth'.
Instead, use the Strategic Friction Protocol. Provide the AI with your thesis statement on, for example, the causes of the Cold War or the characterisation of Lady Macbeth. Then, give it this instruction: "I am aiming for an A*. Do not agree with me. Identify the three weakest points in my argument, provide a compelling counter-argument based on the AQA/OCR mark scheme's focus on nuance, and force me to defend my position without using my original notes."
This creates immediate friction. You are no longer passively reading; you are in a high-stakes intellectual sparring match. By forcing your brain to reconstruct its logic under pressure, you are building the exact mental muscles required for the final 25-mark essay. You can start practising this active debate on platforms designed to challenge, rather than just confirm, your current knowledge.
Phase 2: The 'Synoptic Stress-Test' (GCSE & A-Level Sciences)
The 2025 exam board updates for Biology, Chemistry, and Physics place a heavy emphasis on synoptic assessment. This requires students to link different parts of the specification—for example, explaining how a concept in 'Biological Molecules' relates to 'Transport in Animals'.
Passive revision usually keeps these topics in silos. To create 'Strategic Friction', ask an AI to generate 'Cross-Topic Crucible Questions'. A prompt might look like this: "Create a 6-mark question that requires me to link GCSE Chemistry Topic 2 (Bonding) with Topic 4 (Chemical Changes). Provide no hints. Once I answer, audit my response for technical accuracy and identify where I have used vague terminology."
This forces your brain to retrieve information from two different 'folders' simultaneously. It is difficult, and you will likely get it wrong the first time. That is the point. The 'friction' of failing to find the link in revision ensures you will find it easily in the exam hall.
Phase 3: Stripping the Scaffolding (Mathematics and Problem Solving)
In Maths and Further Maths, students often fall into the 'Worked Example Trap'. They look at a solution, understand it, and believe they can replicate it. To break this, you must use AI to 'un-scaffold' your practice. Many students find that high-quality study materials provide the foundation, but the final leap requires independent struggle.
Instead of asking for a step-by-step solution, ask the AI to:
1. Alter the variables: Change the constraints of a past paper question to make it non-standard.
2. Provide 'Hint Paths' only: Ask the AI, "I am stuck on this integration by parts question. Do not give me the answer. Give me a cryptic clue that points me toward the first step, then wait for my response."
This mimics the role of a tutor while maintaining the 'Desirable Difficulty'. If you get the answer instantly, the neural pathway is weak. If you struggle for ten minutes and then find the solution, that pathway is etched in stone.
The 2025 'Fluency Audit'
To ensure you aren't falling back into passive habits, perform a weekly 'Fluency Audit' on your revision. Ask yourself these three questions:
1. Was I challenged? If you got 100% on your flashcards or practice questions, the material was too easy. You aren't learning; you are performing.
2. Did I use my notes? If you looked at a textbook while answering, that is recognition, not retrieval. Put the notes in another room.
3. Did I synthesise? Did I treat the topic as an isolated fact, or did I force it to connect to another part of the syllabus?
How Thinka Bridges the Gap
While standard AI tools are often designed to be 'helpful' (which usually means 'easy'), Thinka is engineered for academic improvement. By focusing on precision and retrieval, it helps students move past the surface-level summaries that lead to Grade 5s and 6s, and toward the rigorous, friction-heavy practice required for Grade 9s.
Teachers can also benefit from this shift. By using AI to generate practice papers that specifically target 'unseen' contexts and cross-topic links, educators can ensure their students are 'exam-hardened' long before the summer season begins.
Summary: Embrace the Struggle
The goal of revision is not to feel confident; it is to be competent. Confidence is a feeling that often masks a lack of depth. Competence is the ability to retrieve information under the cold, silent pressure of an invigilated exam room.
Next time you use AI in your revision, don't ask it to make your life easier. Ask it to be your toughest examiner. Ask it to find the flaws in your logic, the gaps in your memory, and the weaknesses in your prose. Embrace the strategic friction. In the world of GCSEs and A-Levels, the harder the revision feels, the easier the exam will be.
Related posts
- Jun 4, 2026
The Causality Chain: Mastering the ‘Mechanism of Change’ for Grade 9 and A* Analysis
Stop losing marks for ‘undeveloped’ points. Learn how to use AI as a logical auditor to build the sustained, step-by-step arguments required for top GCSE and A-Level results.
- May 25, 2026
The Semantic Scaffold: Mastering the Technical Register for A* and Grade 9 Success
Stop losing marks to vague phrasing. Learn how to use AI to bridge the 'fluency gap' and master the technical vocabulary required for top GCSE and A-Level grades.
- May 15, 2026
The Mark Scheme Alchemist: Turning Mock Feedback into a Precision Revision Roadmap
Stop letting mock results gather dust. Discover how to use AI to decode examiner feedback and transform mark schemes into a high-precision revision roadmap for GCSE and A-Levels.
- May 5, 2026
The Synoptic Architect: Weaving Cross-Syllabus Connections for A* GCSE and A-Level Success
Master synoptic links to secure top marks. Learn how to use AI to bridge the gap between GCSE and A-Level modules and build the 'integrated understanding' examiners crave.