Artificial Intelligence versus Software Engineers: An Evidence-Based Assessment Focusing on Non-Functional Requirements
Bibliographic record
Abstract
<title>Abstract</title> The automation of Software Engineering (SE) tasks using Artificial Intelligence (AI) is growing, with AI increasingly leveraged for project management, modeling, testing, and development. Notably, ChatGPT, an AI-powered chatbot, has been introduced as a versatile tool for code writing and test plan generation. Despite the excitement around AI's potential to elevate productivity and even replace human roles in software development, solid empirical evidence remains scarce. Normally, a software engineer's solution is evaluated against a variety of non-functional requirements such as performance, efficiency, reusability, and usability, among others. This study presents an empirical exploration of the performance of software engineers versus AI on specific development tasks, using an array of quality parameters. Our aim is to enhance the interplay between humans and machines, increase the trustworthiness of AI methodologies, and identify the best performers for each task. In doing so, it also contributes to refining cooperative or human-in-the-loop workflows in the context of software engineering. The study investigates two distinct scenarios: the analysis of ChatGPT-produced code against developer-created code on Leetcode, and the comparison of automated machine learning (Auto-ML) and manual methods in the creation of a control structure for an Internet of Things (IoT) application. Our findings reveal that while software engineers excel in some scenarios, AI performs better in others. This groundbreaking empirical study helps forge a new pathway for collaborative human-machine intelligence where AI's capabilities can augment human skills in software engineering.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.006 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.000 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".