Hacker News Reader

Comment by jasonv | Hacker News Reader

There was something about the writing in her email that felt.. loaded and questionable without enough factual assertions to carry her point.

In Jeff Dean's email, he made one point that could be pretty salient: she submitted the paper for approval right at deadline. If that's true, that is poorly handled.

The ultimatum is easily a step too far.

And all of this right before a vacation/holiday.

View on HN

I mean the machiavellian interpretation looks like this to me: Google's AI in Ethics initiative is a PR exercise, similar to greenwashing and other models employed by big business. It fundamentally has to be, because large parts of the underlying business model at Google are based on unethical practices. By hiring researchers from underrepresented groups they additionally hoped to a) profit from the power imbalance of the relationship and b) co-opt parts of the social justice movement by having ethics research focus on issues that do not directly touch the bottom line. Now you have a researcher not towing the company line and being fired for it.

I think Google genuinely wants their AI to be better in most of the same ways the ethic researchers want it to be better. I think Google is more willing to compromise on having an ethical AI to meet other goals and priorities than the researchers. Google probably also expects researchers who are publishing papers critical of Google to highlights Google's efforts to mitigate those issues.

> she submitted the paper for approval right at deadline. If that's true, that is poorly handled.

That's not poorly handled on her part. If she'd submitted it late, that would be her problem. If getting a submission right at the last valid time to make a submission (the deadline) is an issue then it's a problem caused by whoever set the deadline, not the submitter.

For internal approval, which has a 2 week process, at the paper deadline. From Jeff Dean's email:

JD> Unfortunately, this particular paper was only shared with a day’s notice before its deadline — we require two weeks for this sort of review — and then instead of awaiting reviewer feedback, it was approved for submission and submitted.

Reply on news.ycombinator.com