Comment by iJohnDoe
5 years ago
The crazy part is that all thousands of employees at Google and they don’t have a single one reviewing what the AI is doing? A review team doesn’t fix mistakes? Even if they do have a team, it’s been proven impossible to get in touch with them to fix anything.
The only team which can fix this issue is the team which trawls social media and occasionally decides to fix an issue, if it becomes enough of a PR problem. Hacker News, Twitter, and Slashdot are apparently the official way to reach humans in Google support.
Of course people review these things. Review teams look at this stuff all the time, and often fix it.
But at Google scale, bad things will happen.
Google serves over five-billion queries every day, and scale for gmail is similar. One-in-a-billion events are absolutely routine.
I don’t think anyone would disagree with your point about the scale at which Google operates or that these type of things happens
However, it’s clear they are not doing enough or not doing anything at all to remedy the problem of having a path to fix things when something does go wrong.
At Google’s scale, they should not be doing automation, AI or otherwise, without support in place when it goes wrong.
"Not doing anything at all" is factually incorrect, and what I was responding to for the gp.
If only there was some way to resolve issues, like having a customer support like with reps who could examine what happened. I hear Google is pretty strapped for cash, though. I have no clue how they'd pay for it, other than, perhaps, taking out a loan on the 11th ball pit or the backup three-story chocolate fountain at Google HQ.