Challenges and Paths Towards AI for Software Engineering

2025-03-28

AI Summary

Current AI applications in software engineering (SE) suffer from narrow task coverage, coarse-grained bottleneck identification, and unclear automation pathways, hindering the realization of highly automated development paradigms. Method: This work introduces the first comprehensive, lifecycle-spanning (requirements, design, testing, operations) structured taxonomy of AI-SE tasks—extending beyond code generation to identify over 40 non-code-centric intervenable tasks. It proposes a unified analytical framework grounded in six recurrent bottlenecks, including contextual insufficiency and weak long-range dependency modeling. Furthermore, it develops four methodological pillars: task-driven knowledge modeling, cross-phase dependency analysis, human-AI collaborative decision-making, and bottleneck-oriented evaluation. Contribution/Results: The study synthesizes 12 prioritized research directions and has been formally adopted as a strategic R&D roadmap by multiple leading SE laboratories.

Technology Category

• Humans and AI: Other Foundations of Human Computation & AI
• Search and Optimization: Metareasoning and Metaheuristics
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

AI for software engineering has made remarkable progress recently, becoming a notable success within generative AI. Despite this, there are still many challenges that need to be addressed before automated software engineering reaches its full potential. It should be possible to reach high levels of automation where humans can focus on the critical decisions of what to build and how to balance difficult tradeoffs while most routine development effort is automated away. Reaching this level of automation will require substantial research and engineering efforts across academia and industry. In this paper, we aim to discuss progress towards this in a threefold manner. First, we provide a structured taxonomy of concrete tasks in AI for software engineering, emphasizing the many other tasks in software engineering beyond code generation and completion. Second, we outline several key bottlenecks that limit current approaches. Finally, we provide an opinionated list of promising research directions toward making progress on these bottlenecks, hoping to inspire future research in this rapidly maturing field.

Problem

• Addressing challenges in AI-driven software engineering automation
• Expanding AI tasks beyond code generation and completion
• Identifying bottlenecks and research directions for AI in software engineering

Innovation

• Structured taxonomy for AI software engineering tasks
• Identifying key bottlenecks in current approaches
• Proposing promising research directions