The increasing adoption of AI-generated code has reshaped modern software development, introducing syntactic and semantic variations in cloned code. Unlike traditional human-written clones, AI-generated clones exhibit systematic syntactic patterns and semantic differences learned from largescale training data. This shift presents new challenges for classical code clone detection (CCD) tools, which have historically been validated primarily on human-authored codebases and optimized to detect syntactic (Type 1-3) and limited semantic clones. Given that AI-generated code can produce both syntactic and complex semantic clones, it is essential to evaluate the effectiveness of classical CCD tools within this new paradigm. In this paper, we systematically evaluate nine widely used CCD tools using GPTCloneBench, a benchmark containing GPT-3-generated clones. To contextualize and validate our results, we further test these detectors on established human-authored benchmarks, BigCloneBench and SemanticCloneBench, to measure differences in performance between traditional and AI-generated clones. Our analysis demonstrates that classical CCD tools, particularly those enhanced by effective normalization techniques, retain considerable effectiveness against AI-generated clones, while some exhibit notable performance variation compared to traditional benchmarks. This paper contributes by (1) evaluating classical CCD tools against AI-generated clones, providing critical insights into their current strengths and limitations; (2) highlighting the role of normalization techniques in improving detection accuracy; and (3) delivering detailed scalability and execution-time analyses to support practical CCD tool selection. The research underscores the continued relevance of classical CCD tools and suggests adopting a hybrid approach that combines both classical and AI-based methods to improve clone detection in the modern era.

Background: Competitive programming platforms such as LeetCode, Codeforces, and HackerRank provide challenges to evaluate programming skills. Technical recruiters frequently utilize these platforms as a criterion for screening resumes. With the recent advent of advanced Large Language Models (LLMs) like ChatGPT, Gemini, and Meta AI, there is a need to assess their problem-solving ability on the programming platforms. Aims: This study aims to assess LLMs’ capability to solve diverse programming challenges across programming platforms with varying difficulty levels, providing insights into their performance in real-time and offline scenarios, comparing them to human programmers, and identifying potential threats to established norms in programming platforms. Method: This study utilized 98 problems from LeetCode and 126 from Codeforces, covering 15 categories and varying difficulty levels. Then, we participated in nine online contests from Codeforces and LeetCode. Finally, two certification tests were attempted on HackerRank to gain insights into LLMs’ real-time performance. Prompts were used to guide LLMs in solving problems, and iterative feedback mechanisms were employed. We also tried to find any possible correlation among the LLMs in different scenarios. Results: LLMs generally achieved higher success rates on LeetCode (e.g., ChatGPT at 71.43%) but faced challenges on Codeforces. While excelling in HackerRank certifications, they struggled in virtual contests, especially on Codeforces. Despite diverse performance trends, ChatGPT consistently performed well across categories, yet all LLMs struggled with harder problems and lower acceptance rates. In LeetCode archive problems, LLMs generally outperformed users in time efficiency and memory usage but exhibited moderate performance in live contests, particularly in harder Codeforces contests compared to humans. Conclusions: While not necessarily a threat, the performance of LLMs on programming platforms is indeed a cause for concern. With the prospect of more efficient models emerging in the future, programming platforms need to address this issue promptly.

With the emergence of Machine Learning, there has been a surge in leveraging its capabilities for problem-solving across various domains. In the code clone realm, the identification of type-4 or semantic clones has emerged as a crucial yet challenging task. Researchers aim to utilize Machine Learning to tackle this challenge, often relying on the Big-CloneBench dataset. However, it’s worth noting that BigCloneBench, originally not designed for semantic clone detection, presents several limitations that hinder its suitability as a comprehensive training dataset for this specific purpose. Furthermore, CLCDSA dataset suffers from a lack of reusable examples aligning with real-world software systems, rendering it inadequate for cross-language clone detection approaches. In this work, we present a comprehensive semantic clone and cross-language clone benchmark, GPTCloneBench1by exploiting SemanticCloneBench and OpenAI’s GPT-3 model. In particular, using code fragments from SemanticCloneBench as sample inputs along with appropriate prompt engineering for GPT-3 model, we generate semantic and cross-language clones for these specific fragments and then conduct a combination of extensive manual analysis, tool-assisted filtering, functionality testing and automated validation in building the benchmark. From 79,928 clone pairs of GPT-3 output, we created a benchmark with 37,149 true semantic clone pairs, 19,288 false semantic pairs(Type-1/Type-2), and 20,770 cross-language clones across four languages (Java, C, C#, and Python). Our benchmark is 15-fold larger than SemanticCloneBench, has more functional code examples for software systems and programming language support than CLCDSA, and overcomes BigCloneBench’s qualities, quantification, and language variety limitations. GPTCloneBench can be found here1.