[mfDjB]

I find the term "ethics" has had its meaning warped lately, ironically for the purposes of "ethics-washing" which is what this article is claiming is happening when big corporations engage in AI ethics. When I worked at a FAANG company in the machine learning field, there was a group of people interested in ethics, but they never seemed to discuss different views on ethics. They often seemed to use the word as a tool to hide behind their collective action to force the company to change course.

For example this article says the following: "Members of Google’s Ethical AI team sent additional demands to Pichai, calling for policy changes, among other things.". If you look at these demands it includes the following:

* The dismissal of the company vice president.

* The reinstatement of Gebru at a higher level than she previously had.

The question as to whether these demands are ethical is not clear at all to me.

Another case in point, anecdotally I saw this a lot at the place I used to work, was the call for regulation to control AI research, and indeed this article talks about this in the section: "Scrap self-regulation". Taken from a different political perspective, someone might argue that giving more control to the state for the purpose of pursuing legitimised violence against AI researchers is unethical (note this is not my view it is an example). I saw people air views like this in the ethics discussions and given very unfavourable treatment, bordering on unethical.

It would be great if AI ethics was a field that produced solutions to problems such as verified datasets, best codes of practice and testing tools. Instead it just seems to be a way to highlight issues and advocate for political positions, which isn't in itself worthless but it feels like the first step in what should be a journey. Unfortunately highlighting problems is really easy, fixing them is the challenging part.

I often got the feeling that the people interested in this just cared about directing the labour of other people for their own interests, e.g. leading from behind rather than leading from the front. All seemed rather unethical to me.

[esja]

In my experience, a surprising proportion of people who get involved in these "ethics" initiatives are mostly there to force their own personal ethics onto the machine, even if it means training the machine to ignore reality and skew it in some direction.

These people are not interested in meta-ethics or any critical examination of their own chosen beliefs. They just want things to be the way they want them to be, and manipulating the AI is a means to that end.

I have no view on whether that is happening at Google. Just a general observation.

[innagadadavida]

Just look at how well it has worked out for the researcher. People who are willing to make a scene and get a lot of attention will win and benefit a lot. If anyone still thinks this is about AI ethics or any other ethics, it definitely is not. This is just getting ahead in a hyper competitive world. Sometimes using questionable means.

[mlthoughts2018]

It reminds me a lot of Robin Hanson’s post “Inequality talk is about grabbing.”

https://www.overcomingbias.com/2013/08/inequality-is-about-g...

[glenneroo]

> The question as to whether these demands are ethical is not clear at all to me.

You left out some vital pieces of context from the same article[0]:

> The note centers on the departure of Google AI ethics researcher Timnit Gebru, which set off protests inside the company. Citing that situation, the employees called for a company vice president, Megan Kacholia, to no longer be part of their reporting chain. “We have lost trust in her as a leader,” the researchers wrote, according to a copy of the later obtained by Bloomberg.

> “Google’s short-sighted decision to fire and retaliate against a core member of the Ethical AI team makes it clear that we need swift and structural changes if this work is to continue, and if the legitimacy of the field as a whole is to persevere,” the letter reads.

> “This research must be able to contest the company’s short-term interests and immediate revenue agendas, as well as to investigate AI that is deployed by Google’s competitors with similar ethical motives,” the researchers added.

As I understand it, the demands themselves are not directly ethical, but in order to do their job (dealing with ethics), they need the VP to stop blocking them in order to make money. As for Gebru receiving a higher position, my take is that, as is with most companies, being a higher-level in the hierarchy allows one to affect more change within the company (or, perhaps in this case, even just to get the job done?).

[0]: https://www.bloomberg.com/amp/news/articles/2020-12-16/googl...

[mlthoughts2018]

Several of the cited passages you link are factually wrong. For example, she definitively resigned and was not fired. Saying she was “fired” instantly reveals partisanship and prejudice to de facto define Google to be in the wrong and the resigned researcher to be in the right completely divorced of the facts.

[itsdrewmiller]

This is misinformation presented very confidently. Unless you give a date for your resignation and the company honors that date, it is a termination. She alluded to one in the future but didn't even give a specific one, as I understand it.

[prepend]

> Unless you give a date for your resignation and the company honors that date, it is a termination

I’m not sure there’s such a fine line. Unless this is defined someone where in Google, I don’t think it’s accurate. I can resign in many ways and then there’s formalities.

[itsdrewmiller]

It is: https://cuiab.ca.gov/board/precedentdecisions/docs/pb39.pdf

[mlthoughts2018]

I think you are asserting misinformation. If Gebru did not list a resignation date and made the resignation conditional on demands in her ultimatum, then the resignation would be effective instantly under the conditions of not fulfilling her ultimatum, just as she said.

You would need example case law of employees declaring a resignation conditional not on any date or time, but on specific demands which are immediately denied by the employer.

>> If you don’t give me a raise then consider me resigned.

>> I don’t give you a raise, therefore I accept your instantaneous resignation.

This linked example where an employee listed a departing date and then was terminated sooner than that does not seen applicable to Gebru’s situation, and I don’t appreciate you saying my comment is misinformation if this is the level of countering material you’re providing. I’m happy to be shown to be wrong, but that’s not what your link describes at all.