How to test with Human-IN-The-Loop - Brijesh Deb
Categories: Podcasts , How To Test This?
The discussion explores the evolution of software testing, from traditional methods to AI-driven approaches, emphasizing critical thinking and responsible AI adoption. It advocates for a “human-in-the-loop” model, stressing governance, skill development, and community engagement to navigate emerging technologies effectively.
How To Test This?
interview episodes where Mamadou N’diaye talks with with software testing experts
- https://podcasters.spotify.com/pod/show/spidey1944
- https://www.youtube.com/@HowToTestThis
- https://www.linkedin.com/in/mamadou-ndiaye-consultant/
Episode Details
- Show Notes: https://podcasters.spotify.com/pod/show/spidey1944/episodes/How-to-test-with-Human-IN-The-Loop---Brijesh-Deb-e3qm9ou
- Published: 2026-10-18T10:17:35Z
- Duration: 00:45:04
- Author: Mamadou N’diaye
Overview
The podcast discusses the evolving role of software testing in the context of emerging technologies, particularly artificial intelligence (AI). The guest shares insights from his extensive career, highlighting the historical development of testing - from early practices without clear distinctions between manual and automated methods, to the impact of Agile, DevOps, and cloud computing. He emphasizes that testing has always involved critical thinking and questioning, and that these core principles remain essential even as the industry adopts new methodologies and tools.
A major focus is on the responsible use of AI in testing. The discussion covers both the potential benefits and significant risks of AI, including bias in training data, over-reliance on AI-generated outputs, and the erosion of human judgment and skills. The guest advocates for a “human-in-the-loop” approach, where AI acts as a co-pilot rather than a replacement for testers. He stresses the importance of evaluating whether AI is truly needed, aligning its use with product strategy, and implementing strong governance at organizational, team, and individual levels. Additionally, the conversation touches on the need for testers to stay current with technology, deepen their understanding of systems, and engage with the broader testing community to continue growing in their roles.
What If
- What if you used AI to stress-test your own judgment as a solo developer?
- Move: Run your existing test cases or product decisions through an AI model (e.g., ChatGPT or Claude), then manually compare its suggestions against your reasoning. Document discrepancies and assess whether they reveal blind spots.
- Why Now?: AI tools are readily accessible, and the risk of over-reliance is increasing - especially for solo operators who lack peer review. Validating AI output against human judgment builds critical discernment before full integration.
- Expected Upside: Sharper decision-making, improved test coverage, and reduced cognitive bias in your development and testing workflows.
- What if you built a personal AI governance checklist for every project you ship?
- Move: Create a lightweight, reusable AI governance template with criteria like: “Was AI used in code/test generation?”, “Was training data contextually appropriate?”, “Was final judgment retained by me?”, and “Is AI use disclosed where necessary?” Apply it before every release.
- Why Now?: Regulatory pressure (e.g., EU AI Act) and market expectations are shifting toward transparency. Solo developers who proactively govern their AI use will avoid reputational and functional risks early.
- Expected Upside: Increased trust in your software, reduced risk of AI-induced failures, and a differentiated personal brand as a responsible builder.
- What if you treated AI as a peer reviewer instead of a generator for your test designs?
- Move: Instead of prompting AI to “write test cases,” feed it your actual test plan and ask: “What edge cases am I missing?” or “Where might bias appear in this logic?” Use it to critique, not create.
- Why Now?: AI-generated content often lacks contextual depth, leading to shallow testing. By reversing the workflow - using AI to review rather than originate - you align with expert advice on keeping humans in the loop.
- Expected Upside: Higher-quality, context-aware test designs with fewer false positives, while maintaining full ownership and accountability as the developer.
Takeaway
- Evaluate whether AI is truly necessary for your testing tasks by analyzing the specific problem you’re trying to solve, avoiding adoption solely due to industry trends.
- Use AI as a feedback tool for your existing test cases by providing context-rich inputs (e.g., geography, domain rules) and asking it to identify gaps or edge cases instead of generating tests from scratch.
- Implement personal AI governance by setting your own guardrails - such as reviewing all AI-generated outputs and disclosing AI assistance - especially when sharing content or making critical release decisions.
- Strengthen your technical foundation by learning enough coding and system architecture to effectively collaborate with developers and design better, more context-aware tests in modern cloud and hybrid environments.
- Engage with professional communities through free meetups or online events to exchange knowledge, reduce isolation, and gain practical insights that improve your testing judgment and career resilience.
For a PDF of longer Software Testing Podcast Episode Summaries with Briefing Notes and more detailed summary notes, visit EvilTester Patreon Podcast Summaries.