In an exploratory pilot, a soft barrier shifted participants from direct copying toward editing, understanding, and iterative resubmission without preventing task completion¶
Claim · All claims
Evidence · 1 study · causalr1·q2
Currency · no dated study recorded · not yet checked against its sources · not reviewed by a person
Subclaims¶
Reading the evidence codes
q is the design tier, i the impact, n the sample, and the kind with its r
the rigour judged within that kind: what each code means.
This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.
q2 i? Group A (barrier) reported less direct copying and more code modification than Group B (4.00 vs. 1.00 for Task 1; 3.67 vs. 1.00 for Task 2), with higher understanding and low frustration. → human-pilot
Evidence¶
human-pilot¶
Iyiola E. Olatunji, Alberick Euraste Djire, Jacques Klein, and Tegawendé F. Bissyandé. 2026. Do Not Copy/Paste: Soft Barriers for Copying in AI-Assisted Programming. Proceedings of the 41st IEEE/ACM International Conference on Automated Software Engineering (ASE '26). https://doi.org/10.1145/3832783.3834554
q2 · i? · causal · r1
Exploratory behavioral pilot with 18 participants using GPT-5.2 in a browser-based tool with two Python tasks; Group A used an invisible-character system prompt, Group B the same model without intervention. Likert responses and clipboard/submission logs were collected; results appear in Table 4.
"Group A reported less direct copying and more code modification than Group B (4.00 vs. 1.00 for Task 1; 3.67 vs. 1.00 for Task 2), consistent with a shift toward editing and reconstruction."
Discussion¶
Related Claims¶
- Diagnostic Review revisions contained larger shares of meaning-change, organization, and content/argument work, while Control revisions remained more dependent on surface-level editing — the general claim this finding bears on, with the studies behind it
- Automatically updating the flowchart and injecting generated code into editors drew participants away from explanations and toward direct code editing — related
- Step-by-step explanations revealed one at a time, plus removed code auto-insertion and copy buttons, encouraged deeper engagement with generated explanations — related
- Output-level Unicode soft barriers can achieve high copy-paste resistance, but effectiveness is highly model- and task-dependent — a broader claim this one bears on
- Constituting AI as a critical file that returns questions instead of revisions is a design requirement for sustaining productive struggle — a broader claim this one bears on
- Pass@1 is not a sufficient success metric for soft barriers because it conflates successful barriers with model generation failures — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
soft-barrier-shifts-users-toward-editing |
title |
In an exploratory pilot, a soft barrier shifted participants from direct copying toward editing, understanding, and iterative resubmission without preventing task completion |
description |
In an exploratory pilot, a soft barrier shifted participants from direct copying toward editing, understanding, and iterative resubmission without preventing task completion |
status |
review |
evidence_strength |
weak |
generated |
by process:wiki-ingest on 2026-10-11 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — human-pilot |
Frontmatter is the page's machine-readable record — see the Schema & Guide.