DOK vs AI: Exploring the Value of Student Learning in the Age of AI

I asked ChatGPT 4.1: "what is the difference between the capacity of students to respond to teacher-assigned tasks and prompts at each DOK Level 1, 2, 3, and 4, and the capacity of AI to respond to user-assigned tasks and prompts?"

I found its response illuminating. My thoughts about this response are below.

Here is the ChatGPT 4.1 response (with zero edits):

---

"This is a rich question—and a timely one! The comparison between student capacity and AI capacity at each DOK (Depth of Knowledge) level reveals a lot about both human learning and the current limits and strengths of AI. Here’s a breakdown for each level, comparing student and AI abilities:

DOK Level 1: Recall & Reproduction

• Students: Can recall facts, define terms, perform simple procedures, or recognize information. This requires attention, memorization, and sometimes pattern recognition.

  • Examples: List the planets, recall a math formula, define “photosynthesis.”

• AI: Excels at this. Can instantly recall facts, definitions, simple procedures, and basic data from vast training or database resources.

  • Examples: Instantly provides lists, definitions, calculations, or matching information.

Difference: Both students and AI can perform well at this level. For students, this is about memory and recognition; for AI, it’s about retrieval from vast data stores. AI generally has a speed and accuracy advantage, but lacks authentic experience or context.

DOK Level 2: Skills & Concepts

• Students: Apply skills, compare/contrast, classify, organize, summarize, and interpret basic relationships. This requires some reasoning and processing, not just recall.

  • Examples: Summarize a story, compare two historical events, use a math concept in a routine problem.

• AI: Performs well with clear instructions. Can organize, summarize, paraphrase, and make basic comparisons with high accuracy. AI can process large amounts of data and identify patterns or relationships.

  • Examples: Summarize an article, compare two scientific theories, solve multi-step math problems.

Difference:Students at this level must apply knowledge, not just recall it. AI can replicate these applications—sometimes even faster or more systematically. The key difference: students may struggle with interpreting the purpose or context of tasks; AI may miss nuance unless the prompt is specific. Students also bring lived experience and emerging understanding, while AI relies on pre-learned associations.

DOK Level 3: Strategic Thinking

• Students: Engage in reasoning, planning, and using evidence. Must justify answers, explain thinking, analyze data, and solve complex problems without a set procedure.

  • Examples: Analyze author’s purpose, develop a scientific hypothesis, solve a real-world math problem.

• AI: Can simulate analysis, make evidence-based arguments, generate hypotheses, and process large data sets. But AI “strategic thinking” is pattern-based and depends on the sophistication of training and prompt clarity.

  • Examples: Analyze themes in a text, generate possible solutions to an engineering challenge, critique an argument.

Difference:Students must construct meaning and may grapple with ambiguity, drawing from experience and adaptive reasoning. AI can simulate strategic reasoning but lacks actual metacognition and personal insight. AI is impressive at pattern recognition and generating plausible analyses, but doesn’t genuinely “think through” problems—its “strategy” is extrapolated from existing models and data.

DOK Level 4: Extended Thinking

• Students: Must synthesize information from multiple sources, conduct investigations, design and execute projects, and reflect on complex, open-ended problems over time. Requires original thought, collaboration, revision, and self-monitoring.

  • Examples: Design a year-long research project, create and refine a business plan, synthesize interdisciplinary concepts.

• AI: Can assist with idea generation, data synthesis, and suggest project structures. Can help manage and process complex information, and simulate project planning. However, AI cannot autonomously sustain inquiry, reflect on its process, or revise based on authentic self-monitoring or emotional feedback.

  • Examples: Propose project ideas, outline research plans, integrate information from multiple disciplines.

Difference:DOK 4 for students is about authentic, sustained, and self-directed inquiry—often with collaboration, iteration, and personal investment. AI can support this level (as a tool), but cannot authentically initiate, sustain, or own a complex project over time. AI lacks agency, motivation, and true reflection. Students bring creativity, resilience, and personal stakes; AI brings efficiency, breadth, and speed but lacks “ownership” and depth of investment.

Summary Table

DOK Level	Student Capacity	AI Capacity	Key Difference
Level 1	Recall, recognize	Recall, retrieve	AI faster/accurate; students build memory
Level 2	Apply, summarize, classify	Summarize, compare, organize	AI is systematic; students interpret nuance
Level 3	Analyze, reason, justify	Simulate analysis, generate solutions	Students construct meaning; AI predicts
Level 4	Synthesize, create, sustain	Assist, suggest, integrate, simulate	Students own/authenticate; AI supports

In essence:

• AI can outperform students at simple recall and procedural tasks (DOK 1, 2), match or exceed at systematic reasoning (DOK 3), and support but not replace authentic, creative, and sustained inquiry at the highest level (DOK 4).

• Students bring experience, emotion, motivation, and meaning-making; AI brings speed, pattern recognition, and breadth but lacks agency, metacognition, and authentic reflection.

If you’d like, I can provide classroom scenarios or prompt examples to illustrate each level for students and AI! Let me know if you want to see those."

---

My personal take on this one-off exchange with ChatGPT 4.1:

When students learn at a DOK1 Level, they are essentially creating the data stores from which they are recalling or reproducing information. This knowledge foundation is critical for higher DOK Levels and, therefore, necessary for human thinking to take place beyond this earliest level (without external support). Yes, AI performs at this level far beyond that of a student because the data stores for AI are the collective knowledge of both the content on which it was trained and its access to all the world's publicly available content via the internet.

The value of human learning at a DOK1 Level to a level of proficiency independent of external (technical) support cannot and should not be replaced by AI in schools if not only on the basis that individuals benefit from forming a foundation of data stores with fundamental knowledge about their world for future higher-capacity thought and application of thought. Of course, what fundamental knowledge ought to be "required" is a separate conversation that should not be conflated with the current focus on the value of human learning at these four DOK Levels.

When we get into DOK2 and DOK3 Levels of thinking, we do so assuming students have formed relevant data stores accessible independent of external supports (i.e. books, peers, internet search engines, AI chatbots) such that a student can exhibit a level of thinking that a school deems necessary in preparation for students to learn outside of a schooling environment. Occasionally, an assessment at these DOK2 and DOK3 levels of learning will be labeled an "open book test" or an instructor might allow students a single notecard with specific knowledge to support higher complexity of application of that knowledge in specific academic domains. Again, we can accept that AI excels at these levels of processing when provided a highly specific prompt with adequate context and purpose. This level of AI application has a great number of useful and relevant applications across industries. The argument is whether the availability of this level of AI ought to replace student capacity for learning at these levels, to which I am openly opposed.

It is worth noting something the AI response asserted: individual humans bring purpose or context at these levels, which suggests that this unique benefit of human learning is distinct from AI at these levels, saying that AI might miss a particular nuance without a grossly explicit prompt and perhaps a few follow-up prompts for further clarification. Unfortunately, a human may just as easily misinterpret or simply forget the purpose or context of a task as an AI chatbot might miss a particular nuance; the lived experience of individuals is not distinct from pre-learned (programmed) associations in AI. The way AI extrapolates from existing models and data is arguably indistinguishable from individuals extrapolating learned strategies from models and data to communicate their mastery of complex ideas. It is the human capacity to learn how to reach these complex levels of thought application through programming via developmentally-appropriate curricula over several years of schooling that enables humans to develop this capacity beyond basic reading, writing, and arithmetic.

Another note worth considering: it is at these DOK2 and DOK3 Levels that schools often define their "passing" level of proficiency for students to receive an accredited degree, with which students use to reflect their assumed mastery of knowledge and process capacity.

It isn't until we explore the DOK4 Level, which is RARELY assessed and MORE RARELY deemed as "passing" (the few exceptions existing as evidence of its vast absence), that we see the value of human capacity for knowledge and process of thought beyond that of AI.

It is at this DOK4 Level that AI becomes the assistant to the brilliance of human thought and ingenuinity. At this level, AI becomes the assistant, able to process huge amounts of knowledge in highly complex processes, allowing the actual "thinking" to happen in the minds of humans beyond anything AI has yet to produce, intentional or not. The use of AI to increase efficiency of this highest DOK level of thinking assumes that the inquiries at this level are rooted in an ever-developing foundation of knowledge and process capacity dependent on an individual human's data stores and capacity for thought at DOK2 and DOK3 Levels that allow for the individual human to evaluate and guide AI resposnes toward responses that an AI, by itself, will never produce independent of its user.

And I only hope the world never sees the day that AI is ever somehow programmed to think autonomously at such a level.

Greg Mullen

July 8, 2025

General Consult

Book Now