As AI tools rush into every corner of software development, one simple question keeps coming back to me. Will AI replace human testers, or will it force us to rethink what great testing even looks like? In today’s conversation, I talk to Santiago Komadina Geffroy, a software engineer at Jalasoft and lecturer at Jala University, about what’s changing, what’s staying, and what teams should do next.
Santiago explains how his daily work and teaching intersect. He points out a hole he often sees. Engineers experiment with large language models without fully understanding how they work, leading to overconfidence and avoidable rework. He advocates clearer interaction patterns between tools and people. Think less about magical cues and more about protocols, context sharing, and agent collaboration. That shift frees testers to do the thinking that AI still struggles with, from exploratory testing and usability assessments to spotting the strange edge cases that only show up when real people use real products.
We also get into prejudice and ethics. AI is only as honest as the data it learns from, and that matters in healthcare, finance and hiring, where a mistake can have life-changing consequences. Santiago calls for stronger education around data quality, authorship, privacy and environmental impact, not as a side note, but as part of the way engineers are trained. He believes that governance helps teams move faster and experience fewer regrets when putting AI into production.
Security is also in the mix. Many AI tools require deep system access. If compromised, they can distort results or leak sensitive information. Santiago is candid about the limits of each individual security. He recommends a culture of shared responsibility, in which engineers understand when to call in security specialists and how to design workflows that keep people informed about consistent decisions.
We conclude with what Jalasoft learned from building with AI within a nearshore model in South America. More thinking time. Smaller, controllable scopes. Clear boundaries between routine automation and human judgment. The headline is simple. AI will change testing. Human testers will remain the core of quality.
Useful links
Subscribe to the Tech Talks daily podcast
#wont #replace #human #testers #Jalasoft