I was reading a book about dry stone bridges, and one of the bridges featured was built across a stream in a protected watershed. The builder said he was surprised the permit for it was approved, he told the client to expect it to be rejected. Talking to the government bureaucrats & scientists, he found out that they had to approve a permit every now and then. Not just because of political reasons, but strategic ones -- if they denied every permit, people would start ignoring and circumventing their department. By approving one every once in a while, they kept themselves in the loop as gatekeepers and prevented more harm than they would have by denying everything. His own bridge project was minimally harmful, and was the token approval so they could keep denying the real problem applications.
I feel like there's a lesson there for other gatekeepers trying to fight the good fight.
It can also mean stealing company IP to try and start a competitor.
Both may be equally forbidden by policy, but they're vastly different acts.
This is my personal anecdote. I had experienced working with an online activist co-worker. She was one of the people who are very popular online, but most of her co-workers hated her. She was extremely entitled, always bossed people (who are at the same level as her) around, did very little of her actual job because she was busy in more important online activism. She had a Messiah Complex, and we were all there just to support her in her worthy cause. That's why I am conflicted when I see activist employees getting fired.
But if a company serially butts heads with people whose actual jobs ARE to hold said company accountable in some way, I find it suspicious.
If a company keeps firing accountants "because they kept yapping about GAAP", wouldn't that raise a red flag with you? Or if a company told their auditor's managers that they'd take their business elsewhere unless more reasonable auditing standards were applied — a practice only all too common in the lead up to financial scandals.
If a president keeps firing inspector generals because apparently they are all part of a Grand Conspiracy against America, wouldn't that be a bit suspicious?
So if Google serially fires AI ethicists, some HN commenters keep insisting that all of them must have been uniquely unreasonable and toxic people, but my money would be on Google having increasing issues with being held to certain ethical standards.
As an aside, paradoxically, it may not be a good sign either if such watchdogs stay at a company forever, appear to be perpetually content and grow highly prosperous. The reason why is left as an exercise to the reader.
Internal audits ranging from a technical issue that impacted clients to sexual harassments investigations are all handled privately and details are shared AFTER the issue is handled. And the details shared are filtered to avoid legal liabilities/bad PR.
An ethics team can certainly coexist with this, the same paper that caused this chain of events could have been shared internally, get a team on board to try to improve on those issues and later publish both the Parrots paper at FAccT and fixes simultaneously in a technical conference.
Still the whole thing is sad because the research they were conducting is impactful