AI made me feel guilty

Over this last couple of months, the Disney project’s implementation has been “pushed” by using AI in all product development phases.

As part of these duties, they assigned us to review feature epics and validate whether acceptance criteria have been satisfied and covered by the codebase that we developed.

At the beginning, I only gave instructions to the AI, and it started to compare the epic against the codebase. However, I was proactive and developed a new skill to automate in some way these revisions and get a detailed report as an output. The process is like this:

1. Analyze the epic (feature), including description, acceptance criteria, user stories, bugs and comments. By using an MCP connected to Jira.
2. Read the codebase and identify previous analysis and create a relationship with classes, functions and variables.
3. Detect discrepancies between the code and acceptance criteria; this will generate a table with risks found including suggestions to solve them.
4. Generate unit tests to “cover” the feature’s behavior.
5. Create a manual testing strategy and verify use cases work properly in a checklist mode. Developers have to do manual tests (QA people usually do it).
6. Generate a report where feature risks are highlighted with code blocks, suggestions for possible solutions, unit testing code, manual testing checklist and conclusions.

Unfortunately the technical leader realized this skill’s potential to be used in all teams. I had a videocall with him and I explained how it works and how to use it. The conversation ended with a moralistic flavor, we deduced this skill will help us a lot to ensure that the feature works correctly. However, this skill can replace the QA workflow and possibly a person.

This small process escalated to the software architects and executives, who gave the green light to be used in all teams. I felt a strange sensation, like waiting for some significant bad news.

A couple of weeks later we heard about a considerable layoff in the project – we never confirmed if that was directly related to the AI skill, but in my head the dates match well – among devs, managers and QA testers (including the QA leader). In that moment I felt so guilty about what I did, not entirely, but I felt I had a big responsibility in that layoff.

I talked about this with my therapist, and I mentioned that I felt like an executioner and guilty about what happened. From his perspective, he told me that I did have a small portion of responsibility. However, my role was to prepare and sharpen the axe. The software architects and executives gave the order as judges. And the technical leader was actually the executioner.

This analogy gave me relief and helped me see past the guilt I’d generated for myself about what happened. If I had refused to do that work, someone else surely would have done it, and perhaps my head would have been cut off in that layoff.

I’ve read some comments saying that AI, in its current state, doesn’t replace employees, but based on what I experienced it’s definitely making significant strides.