The funny thing is that if companies wanted they could probably use AI to instead increase uptime. Keep existing QA teams (instead of replacing them) and then use AI for better and more timely monitoring (and messaging even) and as an additional Always-Testing™ layer of QA

"Just use AI" is definitely not the solve for the problems Github faces.