STRESS 2025: AI-Assisted Programming

28-31 October 2025

Organizers:

  • Steve Bosselmann
  • Daniel Busch
  • Edward Lee
  • Bernhard Steffen

AI-assisted programming is a new trend with high potential. Just ask ChatGPT. It says:

AI-assisted programming considered great for several reasons:

  1. Increased Productivity: AI tools can automate repetitive coding tasks, allowing developers to focus on more complex and creative aspects of programming. This leads to faster development cycles and quicker time-to-market.
  2. Error Reduction: AI can help identify bugs and potential errors in code early in the development process. By providing real-time feedback and suggestions, AI tools can reduce the number of bugs that make it to production.
  3. Code Quality Improvement: AI can suggest best practices and optimize code for performance, readability, and maintainability. It can also enforce coding standards and help refactor existing codebases.
  4. Learning and Skill Enhancement: AI tools can serve as educational resources for developers, offering explanations and examples that help them learn new programming languages, frameworks, or techniques.
  5. Enhanced Collaboration: AI can facilitate better collaboration among team members by providing consistent coding styles and integrating with version control systems to manage code changes efficiently.
  6. Rapid Prototyping: AI can assist in quickly generating prototypes and mockups, enabling developers to experiment with new ideas and iterate on designs more swiftly.
  7. Resource Optimization: AI can help optimize the use of computational resources, such as suggesting more efficient algorithms or identifying unnecessary computations, which can lead to cost savings.
  8. Natural Language Processing (NLP) Integration: With advancements in NLP, AI can understand and process natural language queries, allowing developers to interact with codebases in a more intuitive way, such as generating code snippets from descriptions.
  9. Automated Testing: AI can assist in generating and running test cases, ensuring that code is robust and meets the specified requirements without requiring extensive manual testing.
  10. Accessibility: AI tools can lower the barrier to entry for novice programmers by providing guidance and support, making programming more accessible to a wider audience.

Overall, AI-assisted programming enhances the efficiency, quality, and accessibility of software development, making it an invaluable tool for developers across all levels of expertise.

This sounds wonderful, but what is the reality? Many have experienced also the down-side of AI-support, which really may be quite drastic in cases.

STRESS therefore proposes a hybrid form of AI-assisted programming that exploits technologies developed in decades of formal methods’ research. In particular, it will be shown how to use formal methods, prompt engineering, and software architecture for controlling AI-generated output and reliably combining AI-generated components to a global system. Examples include using domain-specific languages to constrain the context, coordination languages to compartmentalize AI synthesis, and curation of context to improve retrieval augmented generation (RAG). The lectures will provide corresponding guidelines comprisiing

  • tool recommendations
  • success stories
  • warning for pitfalls
  • ways for controlling AI geneated code, e.g., by divide and conquer technology, and
  • best practices.

Besides lectures and tutorials, STRESS will also comprise an AI-asisted programming hackaton where participants are invited to either use their own setup or to profit from the technology proposed during the lectures.