david_iqlabs 3 hours ago

One thing I've noticed with AI generated tests is they can look very convincing even when they're wrong. The output reads confidently but there's not always anything grounding it in real signals.

I've found it works better when the AI is just explaining results that come from deterministic metrics rather than inventing the analysis itself.

Curious how other teams are dealing with that.

  • sriramgonella 3 hours ago

    really good observation. The confidence of the output can sometimes mask the lack of grounding behind it. It almost feels like the emerging pattern is, let AI assist with generation and explanation, but keep the verification layer deterministic and measurable. Curious if you’ve seen teams building internal tooling around that, or if people are mostly relying on existing CI/testing framew

itigges22 3 hours ago

For security vunerability testing on websites I have been making for clients- I almost always hire a senior developer to look over the work and or tests that were created. AI can pass a test, and it can make something that passes a test, but there almost ALWAYS are problems that the senior dev finds with the tests, or with the code that was being tested. Sometimes AI will adjust the code entirely to pass the test or adjust the test to pass failing code.

Another counter-measure I have is to simply lock code before testing. Look over test files, and ensure its not following the happy path.

  • sriramgonella 2 hours ago

    can we even depend on End to end Testing on this AI Tools? but how far these founders can able to rely on that with confidence. I totaly agree for VAPT it will be better