Skip to main content
← How we measure it

Measurement scales

Every survey item we currently administer, grouped by scale, with its response anchors and its internal consistency (Cronbach's α). This is instrument record canon-v1, frozen on 2026-09-22: 86 items. The wording is copied verbatim from the validation exports, so please use it as written.

Download all items as CSV for Qualtrics, Prolific or a spreadsheet.

Rules for these scales

  • No canon item is reverse-scored. A negatively worded construct is named for what it measures and scored in its own direction.
  • Every stem is copied from an export, with its source recorded. An item with no administered stem cannot enter canon.
  • Canon is superseded, never edited. Changes go to canon-v2 with a dated reason.
  • Eight modes describe; the three tier composites are the unit for hypothesis tests.
  • Scales with α below .70 (shown in amber) are short and still being extended. Report them with that caveat, and do not treat a single mode scale as a precise measure of an individual.

AI Interaction Typology

How often a person uses each of the eight modes, plus five scales on how they manage the engagement. Frequency items are answered Never to Always; the engagement scales are 5-point agreement.

The eight modes

Oracle4 items · α = .73
  1. AIT_Oracle1I ask AI to give me a direct answer to a question I'm working on.
  2. AIT_Oracle2I copy or closely paraphrase AI's output into my work without major changes.
  3. AIT_Oracle3I use AI to avoid having to figure something out on my own.
  4. AIT_Oracle4I accept AI's first response as correct without questioning it.

Responses: Never · Rarely · Sometimes · Often · Always

Production Assistant3 items · α = .55
  1. AIT_Production1I use AI to generate a first draft of written work that I then revise.
  2. AIT_Production2I use AI to improve my writing style, grammar, or clarity.
  3. AIT_Production3I use AI to summarize documents, articles, or reference materials.

Responses: Never · Rarely · Sometimes · Often · Always

Tutor3 items · α = .62
  1. AIT_Tutor1I ask AI to explain a concept I don't understand in a different way.
  2. AIT_Tutor2I use AI to create practice exercises or learning materials for myself.
  3. AIT_Tutor3I have back-and-forth conversations with AI to deepen my understanding of a topic.

Responses: Never · Rarely · Sometimes · Often · Always

Collaborative Problem-Solver3 items · α = .71
  1. AIT_CollabPS1I work through a problem step by step with AI, contributing my own reasoning at each stage.
  2. AIT_CollabPS2I use AI as a thinking partner — sharing my approach and asking it to help me refine my logic.
  3. AIT_CollabPS4I go back and forth with AI, building on each other's contributions to solve a complex problem.

Responses: Never · Rarely · Sometimes · Often · Always

Verification Agent3 items · α = .59
  1. AIT_Verify1I check AI's responses against other sources before trusting them.
  2. AIT_Verify2I ask AI follow-up questions to test whether its initial response was accurate.
  3. AIT_Verify3I notice when AI gives me an answer that doesn't seem right and push back on it.

Responses: Never · Rarely · Sometimes · Often · Always

Creative Expander4 items · α = .72
  1. AIT_Create1I use AI to brainstorm ideas or possibilities I wouldn't have thought of on my own.
  2. AIT_Create2I ask AI to help me make unexpected connections between different concepts or topics.
  3. AIT_Create3I use AI to generate multiple alternative approaches to a task before choosing one.
  4. AIT_Create4I prompt AI with "what if" scenarios to explore creative possibilities.

Responses: Never · Rarely · Sometimes · Often · Always

Critical Challenger5 items · α = .81
  1. AIT_Challenge1I ask AI to identify weaknesses or flaws in my arguments or analysis.
  2. AIT_Challenge2I ask AI to present the strongest counterarguments to a position I hold.
  3. AIT_Challenge3I use AI to challenge my own assumptions about a topic.
  4. AIT_Challenge4I ask AI something like "What am I missing?" or "What could be wrong with this reasoning?"
  5. AIT_Challenge5I deliberately ask AI to play devil's advocate on my ideas before finalizing them.

Responses: Never · Rarely · Sometimes · Often · Always

Problem Setter4 items · α = .71
  1. AIT_ProbSet1I use AI to explore different ways of framing or defining a problem before trying to solve it.
  2. AIT_ProbSet2I discuss the context and background of a problem with AI before jumping to solutions.
  3. AIT_ProbSet3I ask AI to help me figure out what the right question to ask is, rather than just answering my first question.
  4. AIT_ProbSet4I use AI to consider a situation from multiple stakeholder perspectives before forming an opinion.

Responses: Never · Rarely · Sometimes · Often · Always

Managing the engagement

AI Calibration4 items · α not yet estimated
  1. AICal_1I can usually tell when AI gives me an incorrect answer
  2. AICal_3I can distinguish between AI responses that are confident vs. actually correct
  3. AICal_4I know which types of tasks AI tends to get wrong
  4. AICal_5I adjust how much I trust AI depending on the topic

Responses: Strongly disagree · Disagree · Neutral · Agree · Strongly agree

AI Metacognition4 items · α not yet estimated
  1. AIMeta_1When using AI, I actively monitor whether I'm thinking critically about the output
  2. AIMeta_3I deliberately check whether I still understand the topic after using AI
  3. AIMeta_4I'm aware of when I'm relying on AI more than I should
  4. AIMeta_5I think about whether using AI is helping or hurting my learning

Responses: Strongly disagree · Disagree · Neutral · Agree · Strongly agree

Cognitive Offloading4 items · α = .80
  1. CogOff_1I don't bother memorizing information that I can easily look up online or ask AI.
  2. CogOff_2When I have access to AI, I spend less time thinking through problems on my own.
  3. CogOff_5If AI is available, I see no reason to work through a problem manually.
  4. CogOff_6I have noticed that I think less deeply about topics since I started using AI regularly.

Responses: Strongly disagree · Disagree · Neutral · Agree · Strongly agree

AI Dependency3 items · α not yet estimated
  1. ADB_1I tend to skip steps in my own thinking process because AI can handle them for me.
  2. ADB_2I often accept AI-generated responses without verifying them against other sources.
  3. DEP_1If AI tools were suddenly unavailable, I would struggle to complete tasks I currently use AI for.

Responses: Strongly disagree · Disagree · Neutral · Agree · Strongly agree

AI Self-Efficacy4 items · α = .85
  1. GSE6AI_1I can always manage to solve difficult problems with the help of AI if I try hard enough.
  2. GSE6AI_2I am confident that I could deal efficiently with unexpected events when using AI tools.
  3. GSE6AI_3Thanks to my resourcefulness, I know how to handle unforeseen situations when working with AI.
  4. GSE6AI_6I can usually handle whatever comes my way when using AI tools.

Responses: Strongly disagree · Disagree · Neutral · Agree · Strongly agree

AI Perceptions Survey

What a person believes about AI across seven dimensions, and whether AI augments or displaces the parts of work they value. All items are 7-point agreement.

Beliefs about AI

Personal Utility4 items · α = .88
  1. APS_PU1AI tools help me accomplish tasks more efficiently.
  2. APS_PU3I benefit personally from using AI in my daily life.
  3. APS_PU6I find AI tools genuinely useful for solving problems I encounter.
  4. APS_PU8My life is easier because of AI technologies.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Societal Benefit4 items · α = .83
  1. APS_SB2Society as a whole benefits from advances in AI technology.
  2. APS_SB4AI can help reduce inequality by making services more accessible.
  3. APS_SB6AI will make education more effective and accessible.
  4. APS_SB8Future generations will be better off because of AI.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Identity Threat4 items · α = .89
  1. APS_IT1AI threatens what makes me valuable as a professional.
  2. APS_IT2The skills I have worked hard to develop are becoming less important because of AI.
  3. APS_IT3AI makes me question whether my expertise still matters.
  4. APS_IT7The rise of AI makes me feel less confident about my professional identity.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Existential Risk4 items · α = .86
  1. APS_ER2The development of AI could lead to outcomes that are catastrophic for society.
  2. APS_ER5AI-powered weapons and surveillance pose a grave threat to human freedom.
  3. APS_ER7Society is not prepared for the risks that advanced AI will bring.
  4. APS_ER8AI could fundamentally destabilize the social and political systems we depend on.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Epistemic Wonder4 items · α = .89
  1. APS_EW2AI inspires me to think about the future in new ways.
  2. APS_EW4Interacting with AI sparks my curiosity about how technology works.
  3. APS_EW5I am excited to see how AI will continue to evolve.
  4. APS_EW7Watching AI develop gives me a sense of wonder about what humans can create.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Ethical Concern4 items · α = .81
  1. APS_EC2AI raises serious questions about privacy that are not being adequately addressed.
  2. APS_EC4The companies developing AI do not give enough consideration to ethical issues.
  3. APS_EC7Current laws and regulations are insufficient to address the ethical challenges of AI.
  4. APS_EC8Society has a moral obligation to slow AI development until ethical safeguards are in place.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Economic Displacement Anxiety4 items · α = .83
  1. APS_EDA1AI will eliminate more jobs than it creates.
  2. APS_EDA5AI will depress wages for most workers.
  3. APS_EDA6I am concerned about the economic security of future generations because of AI.
  4. APS_EDA7The economic benefits of AI will go mostly to large corporations, not ordinary people.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Process valuation

Process Valuation4 items · α = .83
  1. PVF_Val1I find deep satisfaction in working through difficult problems step by step.
  2. PVF_Val2The process of building my skills through practice and repetition is important to me.
  3. PVF_Val3I value the mental effort required to do my work well.
  4. PVF_Val4The time I invest in mastering my craft is a source of personal pride.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Process-Value Augmentation3 items · α = .76

Does AI augment what you value? Behaves like perceived usefulness (latent r .616).

  1. PVF_Fit1AI enhances the parts of my work that I find most meaningful.
  2. PVF_Fit3AI allows me to spend more time on the aspects of my work I value most.
  3. PVF_Fit5AI supports my work without diminishing what makes it personally rewarding.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

Process-Value Displacement3 items · α = .80

Does AI displace what you value? HIGH means more displacement. Not reversed: these are not averaged with PVF_Aug. The only Process-Value scale that relates to Process Valuation (r +.252 as worded). Carries the published Thinker's Paradox findings.

  1. PVF_Fit2AI takes over tasks that I actually enjoy doing myself.
  2. PVF_Fit4AI bypasses the learning process that I consider important.
  3. PVF_Fit6AI removes the challenge from tasks where I find the challenge meaningful.

Responses: Strongly disagree · Disagree · Somewhat disagree · Neither agree nor disagree · Somewhat agree · Agree · Strongly agree

What changes next

  • Author and pilot 2 items each for AIT_Verify, AIT_Tutor, AIT_Production (-> canon-v2).
  • Retire code AIT_CollabPS3; re-author its stem as AIT_Verify6 and test it there.