Using AI as a helper

Modern generative AI tools like ChatGPT are popular because they can access a vast knowledge base while having a very low barrier to entry. A well-written query is all it takes to search through its knowledge. However, we do not get an indication in its answer on how likely the provided information is actually correct. While a web search can also find pages with wrong information, it is easier to assess the quality of that information based on related information, feedback provided to it, credentials of the author, and other details.

A strength of AI is its ability to combine multiple sources of information easily in a way that can answer questions for which no existing document matches all of its aspects, such as "how do I adapt this build file for platform X and runtime environment Y?" or "how do I fix this compilation error in this function?". It can therefore be a useful tool if the result can be validated by you.

However, there is also the risk of getting slightly wrong answers that by being misleading cost more time than what a manual search would take. ChatGPT sometimes even gets very basic queries totally wrong (example Links to an external site.).

Try using a tool like ChatGPT to help you with an aspect of Assignment 3, such as:

• Finding a suitable project.
• Adapting a project to your environment(s), as you may have different environments in your group and potential problems with some of them.
• Troubleshooting configuration, build, or test problems.

After you have tried it, assess the quality of the result and share it on this form.

Note: Do not generate code to include directly in open-source code, as this may violate the license of existing code that the AI used as learning data (by simply scraping content from the web).