In this Agile Alliance Tech Talk, Mark Levison explores the impact of AI on coding practices, critically examining common misconceptions and potential pitfalls. He highlights how AI models do not actually “think” or “reason” but instead generate responses based on statistical probabilities. Levison discusses concerns such as increased duplicated code, reliance on imperfect training data, and diminishing critical thinking skills among developers using AI-assisted coding tools.
Using real-world examples, he illustrates how AI-generated code can introduce inefficiencies and errors, emphasizing the need for developers to approach AI-generated outputs with skepticism. He also explores how AI affects self-confidence in problem-solving and the importance of maintaining human oversight in coding decisions.
To mitigate these challenges, Levison advocates for stronger critical thinking skills, collaboration, and best practices such as test-driven and behavior-driven development. He suggests AI can be beneficial in areas like test automation and bug triage but warns against over-reliance on AI for core development tasks.
The session concludes with a Q&A, where attendees discuss strategies for fostering better coding practices in an AI-driven landscape.
This is an AI-generated summary of the event.
![[Keynote] Panel: Reimagining Agile](https://www.agilealliance.org/wp-content/uploads/2023/08/agile2024-save-the-date-300x158.jpg)
