Post
JA EN

The More You Love the Process, the Harder AI Hits—The 'Helicopter to the Summit' Problem and the Structure of Motivation

Posted
By
13 min read
The More You Love the Process, the Harder AI Hits—The 'Helicopter to the Summit' Problem and the Structure of Motivation
  • Target audience: Software engineers who use AI coding tools daily
  • Prerequisites: Hands-on experience with GitHub Copilot, Cursor, Claude Code, or similar AI tools
  • Reading time: 13 minutes

Overview

Some engineers feel nothing when AI writes their code. Others, with identical skill levels, cannot suppress the urge to write it themselves.

Skill level alone cannot explain this divide. A study by Wu et al. published in Scientific Reports (N=3,562) showed that AI collaboration reduces intrinsic motivation and increases boredom1. Yet the effect does not strike everyone equally. In Anthropic’s study, the 7 participants who maintained strong outcomes under AI assistance used AI not as a “code generator” but as a “dialogue partner for deepening understanding”2.

Here, a variable independent of skill becomes visible. People who find reward in the process itself react to AI in a structurally different way from those who find reward in the outcome. It is like offering a mountaineer a helicopter ride to the summit—the outcome (reaching the top) is the same, but for someone who finds meaning in the climb, it is no longer mountaineering.

This article examines how “attachment to process” interacts with AI adoption, using flow theory and self-determination theory as analytical lenses. To state the conclusion upfront: process orientation is not necessarily an obstacle. If you change how you engage with AI, it can become a strength.

“Process Orientation” vs. “Outcome Orientation”—Two Motivational Structures

The autotelic personality—where the activity is its own purpose

Psychologist Mihaly Csikszentmihalyi introduced the concept of the autotelic personality in his flow theory3. Derived from the Greek auto (self) and telos (purpose), it describes a personality trait of people who engage in activities as ends in themselves.

The characteristics of autotelic personality include3:

  • Motivated by the intrinsic enjoyment of the activity itself
  • Highly curious and persistent
  • Responsive to task-specific incentives rather than external rewards
  • Mastery-oriented rather than performance-oriented

In the software engineering world, this trait is far from rare. JetBrains’ 2025 Developer Ecosystem survey (24,534 respondents) found that 52% of developers write code as a hobby even after coding all day at work4. Coding is both their job and an activity they choose without compensation.

Outcome orientation—when the destination is what matters

On the other hand, some people find reward not in the act of writing code but in the state where the problem is solved and the product works. This can be called outcome orientation.

For outcome-oriented people, coding is one means of reaching a destination. If AI gets them there faster, switching methods carries little psychological resistance.

flowchart TB
    P1["Process-oriented<br>Reward = the act of writing"]
    P1 --> P2["AI writes for you<br>→ Reward disappears"]
    O1["Outcome-oriented<br>Reward = solving problems"]
    O1 --> O2["AI writes for you<br>→ Reward is unaffected"]

    classDef frictionStyle stroke:#cf222e,stroke-width:3px
    classDef harmonyStyle stroke:#2ea44f,stroke-width:3px
    class P2 frictionStyle
    class O2 harmonyStyle

The critical point is that this axis is independent of skill level. A ten-year veteran with an outcome orientation may delegate freely, while a second-year engineer with a process orientation may feel friction with AI.

What AI Takes Away Is Not “Work” but “Reward”

Three threats predicted by self-determination theory

Deci & Ryan’s self-determination theory (SDT) defines three basic psychological needs that sustain intrinsic motivation5:

  1. Autonomy: The sense that you are choosing and controlling your own actions
  2. Competence: The sense that your abilities effectively influence the environment
  3. Relatedness: The sense of connection with others

AI tools threaten autonomy and competence simultaneously.

Regarding autonomy—a workflow where AI generates proposals and humans approve or revise them represents a shift from active creation to passive review. The feeling of “I am choosing” fades.

Regarding competence—when AI instantly produces output equal to or better than skills you spent years acquiring, the sense that “my abilities are making a difference” is destabilized. Mirbabaie et al. termed this phenomenon AI identity threat and identified three predictors: (1) changes in work, (2) loss of status, and (3) perceiving that AI possesses characteristics similar to one’s professional self6.

The motivation decline measured across 3,562 participants

Wu et al.’s 2025 study provided large-scale empirical support for these theoretical predictions1.

MeasureChange in AI collaboration groupEffect size (Cohen’s d)
Intrinsic motivationDecreased-0.29 to -0.51
BoredomIncreased+0.45 to +0.51

The results were consistent across four experiments. Particularly noteworthy was the finding from Study 4: participants who first worked solo and then switched to AI collaboration (Solo→Collab) showed the greatest motivation decline (delta = -0.60) and the greatest increase in boredom (delta = 0.50)1.

The implication runs deep. When people who know the joy of writing code themselves switch to AI, the motivational drop is at its steepest. It is reasonable to infer that the more process-oriented someone is, the larger this drop will be—though the study did not analyze individual differences (personality traits, orientation) as variables.

Less effort, less meaning

Vodiskar & Ruiner’s 2025 study (N=677) revealed another link in this chain7.

AI use had no direct effect on task meaningfulness. However, mental effort functioned as a mediating variable. The non-AI group required greater mental effort, and that effort translated into a stronger sense of task meaning7.

flowchart TB
    A["Without AI"] --> B["Greater mental effort"]
    B --> C["Stronger sense of<br>task meaning"]
    D["With AI"] --> E["Less mental effort"]
    E --> F["Weaker sense of<br>task meaning"]

    classDef effortStyle stroke:#2ea44f,stroke-width:3px
    classDef lossStyle stroke:#cf222e,stroke-width:3px
    class B,C effortStyle
    class E,F lossStyle

Furthermore, Sadeghian et al.’s ACM paper (2024) compared three paradigms of human-AI collaboration and found that the model where humans directly participate and retain accountability for outcomes produces the highest sense of meaning. The supervisory model—where AI oversees the human—produced the lowest8.

In short, “letting AI handle it for efficiency” and “finding meaning in work” exist in a trade-off relationship. Process-oriented people are, in a sense, instinctively resisting this trade-off.

Is Process Orientation a “Weakness”?—Paradoxical Evidence

The conceptual inquiry pattern

Reading this far, process orientation looks like a liability in the AI era. Yet Anthropic’s 2026 study presents a paradoxical finding2.

In an experiment with 52 engineers, the AI-assisted group scored 17% lower on comprehension tests. However, the 7 participants who maintained high scores (a small sample within 52, so generalization requires caution) shared a common trait. They used “conceptual inquiry”—follow-up questions, requests for explanations, concept-level probing—to deepen understanding rather than generate code2.

This group scored above 65%. The full-delegation group scored below 40%.

I interpret this “conceptual inquiry” pattern as a manifestation of process orientation (the original study did not measure process orientation as a variable, so this is the author’s inference). Not “just give me the answer” (outcome-oriented usage) but “explain why it works that way” (process-oriented usage). Concretely, these are the kinds of questions involved:

  • “Walk me through step by step why this code is O(n^2)”
  • “Compare alternative design patterns and their trade-offs”
  • “Explain the root cause of this error along with the debugging steps”

When the way AI was used aligned with the characteristics of process-oriented people, learning outcomes were maximized.

Connection to Amabile’s creativity theory

Harvard Business School’s Teresa Amabile states in her componential theory of creativity9:

People will be most creative when they feel motivated primarily by the interest, enjoyment, satisfaction, and challenge of the work itself—rather than by external motivators.

Among Amabile’s four components of creativity, “intrinsic task motivation” is precisely the core of process orientation. When AI provides instant answers, the wellspring of intrinsic motivation risks running dry. Conversely, when AI is used as a catalyst for generating new questions, process-oriented people may be the ones best positioned to discover creative applications.

Redefining “Craft”

An identity crisis

GitHub Octoverse’s 2025 study documented the identity transformation of the AI era through qualitative research with 22 developers10.

The question developers were grappling with in 2023—“If I’m not the one writing the code, what am I doing?”—is an existential question for process-oriented people. For those whose identity rests on the craft of coding itself, delegating to AI is not merely a tool switch; it demands a redefinition of who they are.

LeadDev’s 2026 article calls this the “AI-driven developer identity crisis”11. Developer identity has been built for years on technical competence, problem-solving ability, and the craft of writing code. AI shakes that foundation.

Four stages of transition

Yet the GitHub Octoverse study also offers hope. Developer AI maturity follows four stages10:

  1. AI Skeptic: Does not trust AI
  2. AI Explorer: Begins experimenting
  3. AI Collaborator: Incorporates AI into daily work
  4. AI Strategist: Redesigns entire ways of working around AI

Notably, developers who reached the Strategist stage described AI adoption not as a loss of craft but as a “reinvention of craft”10.

It is easy to imagine that process-oriented people are more likely to remain at the Skeptic stage. When “writing itself is the purpose,” skepticism toward a tool that takes over that task is natural. But at the Strategist stage, the object of craft has shifted from “writing code” to “defining problems, designing solutions, and ensuring quality.” Those who redefined the target of their process orientation are the ones who overcame the friction.

Redirecting the “Process”

The real question: what process are you attached to?

Synthesizing the analysis so far, the problem is not process orientation itself—it is that the object of attachment is too narrow.

Those attached to the process of coding feel resistance to AI. But those attached to “the process of understanding problems,” “the process of refining designs,” or “the process of deeply exploring user needs” find that AI does not eliminate their process. In fact, by absorbing the implementation workload, AI frees up more time for these higher-order processes.

flowchart TB
    A["Process orientation"] --> B{"Object of attachment"}
    B -->|"The act of<br>writing code"| C["Conflicts with AI<br>Reward disappears"]
    B -->|"Understanding problems<br>and designing solutions"| D["Collaborates with AI<br>Reward increases"]

    classDef frictionStyle stroke:#cf222e,stroke-width:3px
    classDef harmonyStyle stroke:#2ea44f,stroke-width:3px
    class C frictionStyle
    class D harmonyStyle

The 7 high-scoring participants in Anthropic’s study had made this shift in practice2. They treated their AI dialogue not as “delegating code generation” but as “exploring concepts.” They directed the traits of process orientation—curiosity, persistence, mastery drive—toward their interaction with AI.

Not “climbing” but “exploring”

Return to the mountaineering metaphor from the beginning. If a helicopter carries you to the summit, a mountaineer would not call it mountaineering. But if the purpose is “exploring unknown mountain terrain,” the helicopter is just transportation to base camp—the exploration itself remains.

AI may skip the “climbing” of coding. But the “exploration” of software engineering does not disappear. Tackling unknown problems, iterating on system design, refining user experience—all of this remains. The key to adaptation for process-oriented people is not asking “what to climb” but “what to explore.”

Conclusion

When process-oriented people struggle with AI, it is not a character flaw. It is a structural problem where the source of reward is being displaced by AI.

Wu et al.’s study showed that AI collaboration reduces intrinsic motivation1, and Vodiskar & Ruiner’s study revealed the mechanism by which reduced mental effort weakens the sense of meaning7. Within the self-determination theory framework, AI threatens the two basic needs of autonomy and competence simultaneously5.

Yet the “conceptual inquiry” pattern identified in Anthropic’s study illuminates a path for process-oriented people to harness AI2. The “reinvention of craft” documented by GitHub Octoverse points in the same direction10.

Process orientation can be either a weakness or a strength. The fork in the road lies in whether you can update the “process” you are attached to. From the process of writing code to the process of understanding problems and designing solutions. The intellectual activities that AI cannot replace are still abundant.

References

References corresponding to citation numbers in the article are listed in numerical order.

Additional references (not cited by number in the article)

On the accuracy of citations: The studies cited in this article have been verified through academic databases (Google Scholar, PubMed, ACM Digital Library, etc.), official journal websites, and cross-referencing with multiple independent sources. While full-text access may be restricted for some papers, abstracts, DOIs, and key findings have been verified through official academic databases and reliable sources.

  1. Human-generative AI collaboration enhances task performance but undermines human’s intrinsic motivation - Wu, Liu, Ruan, Chen, Xie, Scientific Reports / HBR version (2025). 【信頼性: 高】 Peer-reviewed journal paper. 4 experiments, N=3,562. ↩︎ ↩︎2 ↩︎3 ↩︎4

  2. How AI Assistance Impacts the Formation of Coding Skills - Shen, Tamkin, Anthropic (2026). arXiv: 2601.20245. 【信頼性: 高】 52 engineers. RCT. ↩︎ ↩︎2 ↩︎3 ↩︎4 ↩︎5

  3. Csikszentmihalyi, M. (1990). Flow: The Psychology of Optimal Experience. Harper & Row. Systematic treatment of autotelic personality: Baumann, N. (2021). “Autotelic Personality.” In Advances in Flow Research (2nd ed.), pp. 231-261. Springer. 【信頼性: 高】 ↩︎ ↩︎2

  4. State of Developer Ecosystem 2025 - JetBrains (2025). 【信頼性: 中〜高】 Large-scale developer survey: 24,534 respondents, 194 countries. ↩︎

  5. The “What” and “Why” of Goal Pursuits: Human Needs and the Self-Determination of Behavior - Deci, Ryan, Psychological Inquiry (2000). / Gagné, Deci, “Self-determination theory and work motivation,” Journal of Organizational Behavior (2005). 【信頼性: 高】 Established psychological theory. ↩︎ ↩︎2

  6. The Rise of Artificial Intelligence – Understanding the AI Identity Threat at the Workplace - Mirbabaie, Brunker, Mollmann Frick, Stieglitz, Electronic Markets (2022). 【信頼性: 高】 Peer-reviewed journal paper. ↩︎

  7. Collaboration between individuals and AI: fusing mental effort and AI for work meaningfulness - Vodiskar, Ruiner, AI & Society (2025). 【信頼性: 高】 Peer-reviewed journal paper. N=677. ↩︎ ↩︎2 ↩︎3

  8. The Soul of Work: Evaluation of Job Meaningfulness and Accountability in Human-AI Collaboration - Sadeghian, Uhde, Hassenzahl, Proceedings of the ACM on Human-Computer Interaction, CSCW1 (2024). 【信頼性: 高】 Peer-reviewed ACM paper. ↩︎

  9. Componential Theory of Creativity - Amabile, Harvard Business School Working Paper 12-096 (2012). 【信頼性: 高】 Established creativity theory. ↩︎

  10. The New Identity of a Developer - Kalliamvakou, GitHub (2025). 【信頼性: 中〜高】 Qualitative study of 22 developers. ↩︎ ↩︎2 ↩︎3 ↩︎4

  11. Managing the AI-driven Developer Identity Crisis - McMahon, LeadDev (2026). 【信頼性: 中】 Practitioner-oriented media article. ↩︎

AI・テクノロジー
AI Motivation Flow Self-Determination-Theory Identity Cognitive-Science
This post is licensed under CC BY 4.0 by the author.