Skip to main content

Responsibility Without Authorship

25-minute Talk

When AI writes the code, who is responsible? Testers have always been accountable for systems they didn't build. That's not a limitation. It's the defining skill of the AI era.

Virtual Pass session

Timetable

11:45 a.m. – 12:30 p.m. Tuesday 17th

Room

Room F2 - Track 2: Talks

Artificial Intelligence (AI) Career Development Collaboration & Communication

Audience

Testers, Test Leads, anyone questioning their role in an AI-influenced world

Key-Learnings

  • AI is great at plausible output and terrible at sensing hidden risk. That gap has always been home turf for testers.
  • The center of gravity is shifting from authorship to stewardship: communication, systems understanding, and judgment matter most.
  • You are not being displaced by AI. You are being promoted by it. The shift lands right in your lap.

The skill you've been building your whole career (feeling accountable for systems you didn't build) might be the most important capability in modern software development right now.

When AI writes the code, who is responsible for what it does?

This isn't hypothetical anymore. Developers are generating code they haven't written line by line. Teams are shipping features built by tools that can't explain their own reasoning. And the old model of "I wrote it, so I own it" is quietly falling apart.

But as testers we've never had that luxury. We've always been accountable for systems we didn't craft ourselves. We interrogate, we model, we evaluate consequences, and we remain responsible anyway. That's not a limitation of our role. That's the skill the entire industry now needs.

AI tools are very good at plausible, mostly-working output. They are terrible at knowing when something is unacceptable, sensing hidden risk, or respecting constraints that were never written down. That's where we live. That has always been our home turf.

Here's what that looks like. I spent three mornings debugging a performance problem in a system I'd built with AI assistance. The AI proposed a timing correlation that looked convincing. But something didn't fit. I kept asking until we found it. I noticed the numbers were wrong, asked why, and refused to accept "it should be fine" as an answer. That's testing. AI didn't invent that skill. It made it indispensable.

I don't think testers are being displaced by AI. I think we're being promoted by it. The center of gravity is shifting from construction to judgment, from authorship to stewardship. And that shift lands right in our lap.

Related Sessions

There are currently no related sessions listed. Please check back once the program is officially released.