cs.SEAug 23, 2026

Do Not Copy/Paste: Soft Barriers for Copying in AI-Assisted Programming

Authors: Iyiola E. Olatunji, Alberick Euraste Djire, Jacques Klein, Tegawendé F. Bissyandé

Organizations: University of Luxembourg Luxembourg, Luxembourg

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Copying a function from a chat window into an editor takes less than a second. For many uses of AI coding tools, that speed is the point; in settings such as programming education, code review, and security-sensitive development, it can also be the problem. This paper frames copy-paste as an AI code handoff problem: the moment model-generated text crosses from a conversational context into executable or committed software is a design boundary that current tools leave largely unmanaged. We argue that AI coding assistants should not only be evaluated by the code they generate, but also by how they mediate the transfer of that code into software artifacts. We propose soft barriers as one class of handoff-aware mechanisms. Soft barriers preserve access to AI assistance while making unexamined transfer less frictionless. As an initial technical probe, we instantiate this idea using Unicode output perturbations that preserve visual readability but disrupt naive copy-paste execution. We introduce Copy-Paste Resistance (CPR), the fraction of functionally correct clean solutions that become syntactically invalid after perturbation. Across HumanEval and MBPP with four LLMs and four perturbation families, we find that output-level barriers can achieve high copy-paste resistance, but their effectiveness is highly model- and task-dependent. An exploratory pilot with 18 participants provides early evidence that soft barriers can shift users from direct transfer toward editing and reconstruction. We do not present Unicode perturbations as a deployment-ready solution; rather, we use them as a minimal probe for a broader research agenda on practical, transparent, and policy-aware AI code handoff.

Figures & tables

Handoff typeMechanism
DirectCode transfers immediately with no intervention
Explanation-firstAssistant provides reasoning before code; promotes inspection
Soft-barrierCode is readable but naive copy-paste execution fails
Review-gatedTransfer requires a summary, test, or confirmation step
Test-gatedInsertion is blocked until the learner runs or writes tests
Provenance-markedGenerated code carries metadata for attribution or auditing
Table 1. Handoff design space for AI coding assistants.
Figure 1. Example of a homoglyph intervention. The generated code appears visually valid, but pasting it into an editor triggers an execution error.
FamilyMechanism
Invisible charsZero-width characters (U+200B, U+200C, U+200D) inside identifiers
HomoglyphsASCII letters replaced by visually identical Unicode equivalents
BIDI reorderingDirectionality overrides (U+202E, U+202D) altering internal byte order
DeletionsBackspace/delete controls (U+0008) erasing adjacent rendered characters
Table 2. Unicode perturbation families.
Figure 2. Pass@1 on HumanEval and MBPP under each intervention and baseline.
DatasetInterventionClaudeDeepSeekGPT-5.2Codex
HumanEvalDeletion0.8460.0400.3500.330
Homoglyph0.0080.6400.1960.130
Invisible chars0.7660.2480.8460.773
Reordering0.9430.0640.3160.373
MBPPDeletion0.9820.0110.0100.090
Homoglyph0.9860.7960.0000.005
Table 3. CPR scores. Higher values indicate stronger handoff resistance. Values in bold exceed 0.75.
 Task 1Task 2
Metrics | GroupsABAB
Copied directly ↓\downarrow3.835.003.175.00
Modified code ↑\uparrow4.001.003.671.00
Understood code ↑\uparrow4.503.004.002.00
Felt frustrated ↓\downarrow1.671.001.831.00
Helped learn ↑\uparrow4.002.835.003.50
Table 4. Human experiment results. Arrows indicate the direction favorable for learning. No. of retries denotes the average number of submission attempts before completing task.
More in this topic

Human-AI Collaboration

Human-AI collaboration studies how people and AI systems divide, coordinate, and adapt work to support decisions, creativity, and learning.

See all 446 papers

Explore similar work

CardsList
Sep 15, 2026

Does AI Assistance Leave a Temporal Fingerprint? Detecting Overreliance in AI-Assisted Writing and Programming

The rapid adoption of generative AI has made final artifacts unreliable evidence of student learning, and AI detectors that examine only the finished product are inaccurate and ethically contentious. Process data offers an alternative, but prior work covers only English essay writing. We ask whether AI assistance carries a temporal signature, whether it generalizes from writing to programming, and whether it distinguishes ordinary collaboration from wholesale delegation. We analyze three public corpora: CoAuthor (1,447 keystroke-level co-writing sessions), RealHumanEval (editor telemetry from 243 programmer records), and a pre-LLM CS1 corpus (5.1 million keystrokes) as a human-only baseline, comparing minimal-AI work, collaborative AI use, and simulated wholesale delegation. Three findings emerge. First, the signature generalizes: AI contributions arrive in bursts far outside the author's own baseline in both mediums (paired d_z = 1.13 and 3.54). Second, engagement diverges by medium: 93% of AI-inserted characters survived to writers' final documents, while only 14% of accepted code suggestions survived intact. Third, classifiers using only observable temporal features separate simulated delegation from authentic work nearly perfectly (F1 ≥\geq 0.997; at most 0.5% of real work misclassified), while ordinary collaboration remains hard to distinguish from unassisted work. Temporal evidence flags wholesale delegation rather than assistance, positioning process visibility as a candidate evidentiary basis for academic integrity, pending validation in authentic coursework.

Eduardo Davalos, Yike Zhang

Trinity University, San Antonio, TX 78212, USA · St. Mary's University, San Antonio, TX 78228, USA

Generative AI in EducationAcademic IntegrityHuman-AI Collaboration
AskPDF
Sep 11, 2026

What is the Difference Between Me and You? Benchmarking the Quality Gap Between Human-Written and AI-Generated Code

AI coding assistants are becoming co-authors of production software, yet their evaluation centers on functional correctness, leaving open whether their code differs from human code in the quality dimensions dominating lifecycle cost. We compare human-written and AI-generated code at scale: 787,562 function pairs across Python, Java, and C, each human function mined from open-source repositories paired with implementations generated from its docstring by three AI assistants (OpenAI GPT models, DeepSeek-Coder, Qwen2.5-Coder). We characterize structural complexity and statistical naturalness, and map static-analysis findings onto Orthogonal Defect Classification for defects and the Common Weakness Enumeration for vulnerabilities, making authors and languages directly comparable. AI-generated code is structurally compressed and stylistically templated: roughly half the size and branching of human code, clustering apart at the style level. Defect profiles differ in kind: human code concentrates issues of mature codebases, AI code repetitive boilerplate; security is language-dependent, with LLMs producing more, and more severe, findings in Python and Java but fewer high-severity memory-safety findings than humans in C. Once size is controlled for, complexity metrics carry little signal, while naturalness separates authors. Finally, we release CQBench, a benchmark of 27,346 issue-prone tasks with baselines and an evaluation pipeline for quality assurance and security testing.

Cristina Improta, Pietro Liguori, Domenico Cotroneo

University of Naples Federico II, Naples, Italy · University of North Carolina at Charlotte, Charlotte, NC, USA

LLM SecurityCode Generation EvaluationCode Generation Benchmarks
AskPDF
Apr 20, 2026

Scaling Human-AI Coding Collaboration Requires a Governable Consensus Layer

Vibe coding turns natural-language intent into working software, but code generation alone does not give humans and agents a shared account of what they are building. As collaboration spans participants, components, and releases, design choices must remain understandable and open to revision. We argue that scaling human and AI coding collaboration requires a governable consensus layer. Agentic Consensus organizes collaboration around a versioned project-state model linking selected intent, structural commitments, executable artifacts, and evidence. Participants can clarify intent before implementation and coordinate proposed changes against commitments that humans can inspect and revise. Bidirectional synchronization connects intended changes to implementation and exposes conflicts or stale evidence as artifacts evolve. We implemented its review workflow in Consensus Control. In a small exploratory human study, 10 participants used the system on projects from their everyday work and reported clearer, faster understanding of agent-submitted changes. A three-condition comparison with the same participants collected 60 responses, with near-ceiling decision scores and measured representation and pipeline costs. Our research agenda asks how shared project state can preserve human understanding and control as agent contributions accumulate.

Tianfu Wang, Zhezheng Hao, Yin Wu +7

HKUST (GZ) Guangzhou, China · ZJU Hangzhou, China · USTC Hefei, China +1

AI Agent BenchmarksAI Agent GovernanceKnowledge GraphsHuman-AI Collaboration
AskPDF