Why don’t we replace the testers with AI? Postscript

Publication date Read time 1 min Tags testing

When researching my previous article, I stumbled upon an essay by Dave Karpf titled “Two Failure Modes of Emerging Technologies”.

The first failure mode is when the technology is actually successful, finds itself used outside of planned context and that causes a whole swath of unintended consequences. Second - technology fails to deliver what it promised, but still finds itself widely used and incorporated into critical systems. Everyone learns to works around problems, or worse - stops to see there are problems to begin with.

Most of AI critique falls closer to the first mode of failure. My article is no different - my main argument is that AI will not get as good as it needs to be to replace testers anytime soon.

But if I were taking bets, my money would be on AI following the second mode of failure - it will not live up to it’s hype, and still get widely deployed.

It’s unfortunate I did not spend more time reflecting on this.