The Adoption and Impact of AI-Powered Coding Tools Among Swedish Software Developers

AI-powered coding tools such as GitHub Copilot, ChatGPT, and Amazon CodeWhisperer are increasingly integrated into software development workflows, offering automated code generation and support. However, their adoption raises concerns about trust, ethical use, and organizational policies, especially in contexts with evolving AI regulations like Sweden. This thesis investigates how Swedish software developers trust and use AI-generated code across tasks and tools, and what ethical issues they associate with these technologies. A structured survey was conducted with 83 software professionals in Sweden, collecting data on usage patterns, trust levels, perceived benefits, and ethical concerns. The results show that while developers widely use AI tools for simple tasks like boilerplate code, their trust declines significantly for complex or security-critical code. Nearly all respondents review AI-generated code before implementation, and many express concerns about licensing violations, bias, and long-term skill erosion. These findings highlight the need for clearer guidelines, transparency in AI assistance, and stronger organizational policies to ensure responsible adoption.