CodeEvolve: Automating Legacy App Migration; The Battle over the Regulation of Artificial Intelligence

How can AI be created, trained, and used more responsibly in the United States?
Many applications run on outdated frameworks, meaning lagging in security and performance. CodeEvolve is a desktop application leveraging large language models (LLMs) to automate the migration of legacy front-end applications to modern frameworks. PySide6 is used to create modular widgets, including a terminal and multiple selection options. A dependency parser and regex scanner first read through the legacy app files, then a new NextJS project is generated that mimics the structure of the legacy app. Finally, the project is filled with translated code from API calls to an LLM. The application struggles to properly modernize the structure of complex legacy apps, but minimal development is necessary after migration for simple apps. Future work should be done on the number of frameworks that can be migrated, since currently only Angular and Java Server Pages (JSP) are available.
Since 2010, how have tech companies, safety advocates, critics of algorithmic bias, fair labor proponents, defenders of intellectual property, and others competed to determine the standards governing AI systems’ development and use? AI companies warn that legal and administrative battles may deprive the US of its leadership in AI. Many artists, musicians, and other creatives contend that AI violates their intellectual property rights. Unions and some thinktanks resist AI, arguing that it may accelerate automation and consequent unemployment. Civil rights advocacies strive to stop algorithmic bias. Divisions among both companies and nonprofits limit their efficacy.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Briana Morris

STS Advisor: Peter Norton

Technical Team Members: