Coding tools powered by artificial intelligence (AI) can offer a huge productivity boost to developers, so much so that many software companies are already encouraging their teams to adopt these tools. But is the code created by AI assistants any good? Is it secure? Can it be trusted? These are some of the fundamental questions currently being asked as development teams continue to weigh the pros and cons of AI-assisted coding.
To understand the quality of AI-generated code better, we asked the popular AI assistant, ChatGPT, to generate code from a simple prompt, then ran static code analysis on the code to understand what kind of quality issues and security vulnerabilities might emerge. In this talk, we will share our findings and answer the questions:
Can AI-generated code be trusted?
How can we make the most of emerging AI tools without compromising code quality and security?