TestGuild Automation Podcast Testing AI's Promises with Michael Bolton
May 4, 2025
Michael Bolton, a legendary software testing expert, joins the discussion to dissect the realities of artificial intelligence in testing. He challenges the hype surrounding AI, describing it as a collection of probabilistic text generators that can mislead with overconfidence. Bolton offers insights on when AI can enhance testing and the risks involved in its blind adoption. He emphasizes the vital role testers will play in an AI-driven future, advocating for critical thinking and responsible integration of these technologies into workflows.
AI Snips
Chapters
Transcript
Episode notes
AI's Illusion of Intelligence
- AI tools are probabilistic text generators, not truly intelligent entities.
- They produce confident outputs that can mislead users into overtrusting them.
Use AI in Low-Risk Contexts
- Use AI when the risk is low and the output volume is manageable for human scrutiny.
- Avoid using AI for critical decisions without expert oversight to prevent costly mistakes.
Use AI as Inquiry Tool Only
- Use AI output as a tool for inquiry and discovery, not as a decision-maker.
- Always scrutinize AI results, especially in high-risk scenarios or when variability matters.