Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google is investigating the actions of another top AI ethicist

axios.com

Post reply on HN
Print view View original

101–110 of 125 posts

Re: Google is investigating the actions of another top AI ethicist

#101

Post by jsinai »

visarga wrote: post #67
NotPavlovsDog wrote: post #46

Earlier quoted context omitted.

A large part of them believe that a researcher has to be an activist, i.e. research=activism. Many recent textbooks on social science see the following as credible resarch: "Emancipatory research: Research that exposes underlying ideologies in order to liberate those oppressed by them." (Zina O'Leary textbook, used by many schools, over 2K citations for multiple editions) So if a corporate behemoth such as Google wan…

It was clearly a mistake. They don't act in a constructive way. Let's take Timnit's paper for example. She found out that Google had been using a biased language model in search. The bias was like - 'male' association with 'doctor' as opposed to 'woman' association to 'nurse'. But she didn't show how this is harming anyone in a concrete way. Just theoretical. And then she offered no solution, just blaming the work of…

> But she didn’t show how this is harming anyone in a concrete way.

To roll with the example “man” ~ “doctor”, “woman” ~ “nurse”, the harm is having a giant and widely used search engine reinforce baseless gender biases, ie that there is no underlying reason why women should be nurses and men doctors. What is the harm you may ask? The harm may be subtle, eg being surprised when you find out your next doctor is a woman or your next is a man. It could suppress career choices and aspirations, and it could even be financial, eg reinforcing systemic pay gaps.

Re: Google is investigating the actions of another top AI ethicist

#102

Post by dragonwriter »

knorker wrote: post #83
dragonwriter wrote: post #69

Earlier quoted context omitted.

> The (1) case is a bit inaccurate/misleading. No, its not. > Gebru tweeted the name of the employee Sure, more to the point Gebru tweeted that Mitchell’s corporate email appeared to be nonfunctional, sure. > Axios then reached out to Google That seems likely to be the sequence of events, sure. Usually and ethically, a company that was in exactly the circumstances Google described would have: (1) Declined comment, or…

The only way I can read your argument is if I were in the spirit of "the big guy is always wrong". If you're accused of a SECOND action taken against an ethicist, and AGAIN it's not because of anything you did that was bad, then yes your hand is actually being forced to say more than "no comment". It's absolutely incredible how much Google has not spoken out to defend itself against the lies upon lies upon inconsiste…

> The only way I can read your argument is if I were in the spirit of "the big guy is always wrong".

How about “the guy burning the person they are currently in a business relationship with without getting all the facts is wrong”. Or “a wrong by party A against party B does not justify a wrong by B against C.”

> It's absolutely incredible how much Google has not spoken out to defend itself against the lies upon lies upon inconsistencies that Timnit has thrown out. And now another case pops up?

Er, nothing in Google’s story, even taken as gospel truth, indicates that either Gebru’s fact claims in this case were lies or that her speculations were unreasonable or inconsitent in her position given the observable facts. So, your characterization seems...misplaced, at best, even if your characterization of her past actions was accurate. (And in the cases where Google has presented contrary stories to Gebru on other points, Google’s own stories have been outright self-contradictory whereas Gebru’s were at least internally consistent, so I can either trust Gebru or neither.)

Re: Google is investigating the actions of another top AI ethicist

#103

Post by mikejb »

dragonwriter wrote: post #92
throwawaysea wrote: post #87

Earlier quoted context omitted.

Isn’t that a one sided expectation though? Timnit Gebru publicly tweeted about this second activist losing email access. By calling Google out in such a public manner, I feel like Google is forced to respond publicly, to set the record and prevent the early viral spread of these activists’ one sided take/misinformation. Otherwise what happens is journalists like Kara Swisher source entire stories from these activists…

> Isn’t that a one sided expectation though? Yes, Google’s ethical responsibility in a current employer/employee relationship with Mitchell is different than Gebru’s ethical obligation to her former employer with whom she is already in a public, contentious battle. And even if the ethical obligations were identical, Gebru’s violation toward Google wouldn’t excuse Google’s toward Mitchell.

> And even if the ethical obligations were identical, Gebru’s violation toward Google wouldn’t excuse Google’s toward Mitchell.

What about Mitchell's violation towards Google?

Re: Google is investigating the actions of another top AI ethicist

#104

Post by gillesjacobs »

visarga wrote: post #50
nostromo wrote: post #14

Earlier quoted context omitted.

Many of the people in this field aren’t ethicists, they’re activists.

True, they only support their own political faction under the guise of "identity politics", not all people. For example Andrew Ng is considered "white adjacent" because he's the only non-white in an article about the history of AI. Asians are not favored with this group of activists.

Racialists on either side of the aisle really enjoy their mental gymnastics.

Re: Google is investigating the actions of another top AI ethicist

#105

Post by hnlurker »

visarga wrote: post #67
NotPavlovsDog wrote: post #46

Earlier quoted context omitted.

A large part of them believe that a researcher has to be an activist, i.e. research=activism. Many recent textbooks on social science see the following as credible resarch: "Emancipatory research: Research that exposes underlying ideologies in order to liberate those oppressed by them." (Zina O'Leary textbook, used by many schools, over 2K citations for multiple editions) So if a corporate behemoth such as Google wan…

It was clearly a mistake. They don't act in a constructive way. Let's take Timnit's paper for example. She found out that Google had been using a biased language model in search. The bias was like - 'male' association with 'doctor' as opposed to 'woman' association to 'nurse'. But she didn't show how this is harming anyone in a concrete way. Just theoretical. And then she offered no solution, just blaming the work of…

I have no ML background, I’m not trying to be overly critical, but why would the association go in that direction? Wouldn’t the model be the other way? Ie doctor is male and nurse female.

Also, it’s a pretty shitty model since females dominate the medical field in both real and per capita numbers.

Re: Google is investigating the actions of another top AI ethicist

#106

Post by dragonwriter »

mikejb wrote: post #103
dragonwriter wrote: post #92

Earlier quoted context omitted.

> Isn’t that a one sided expectation though? Yes, Google’s ethical responsibility in a current employer/employee relationship with Mitchell is different than Gebru’s ethical obligation to her former employer with whom she is already in a public, contentious battle. And even if the ethical obligations were identical, Gebru’s violation toward Google wouldn’t excuse Google’s toward Mitchell.

> And even if the ethical obligations were identical, Gebru’s violation toward Google wouldn’t excuse Google’s toward Mitchell. What about Mitchell's violation towards Google?

> What about Mitchell's violation towards Google?

If we accept Google's own claims, they have an automated indication which leads to suspicion of that and on ongoing investigation, not even something where they are prepared to claim an actual violation. i.e., exactly the circumstances where every half competent organization would decline comment (potentially citing “personnel matters” until they'd actually completed an investigation.)

Re: Google is investigating the actions of another top AI ethicist

#107

Post by avereveard »

nostromo wrote: post #14
esja wrote: post #8

“In this instance, yesterday our systems detected that an account had exfiltrated thousands of files and shared them with multiple external accounts.” If true, that is damning, and would demonstrate once again that being an “ethics researcher” does not mean that you are are any more ethical than the average person. It just means you are more interested in the subject. As a side note, I wish this field was more intere…

Many of the people in this field aren’t ethicists, they’re activists.

There's plenty of both, but Google specifically hired a majority of activist for this panel

Re: Google is investigating the actions of another top AI ethicist

#108

Post by mikejb »

dragonwriter wrote: post #106
mikejb wrote: post #103

Earlier quoted context omitted.

> And even if the ethical obligations were identical, Gebru’s violation toward Google wouldn’t excuse Google’s toward Mitchell. What about Mitchell's violation towards Google?

> What about Mitchell's violation towards Google? If we accept Google's own claims, they have an automated indication which leads to suspicion of that and on ongoing investigation, not even something where they are prepared to claim an actual violation. i.e., exactly the circumstances where every half competent organization would decline comment (potentially citing “personnel matters” until they'd actually completed…

So you say you suspect they have an ongoing investigation, and that would be "circumstances where every half competent organization would decline comment"

What substantiates this conclusion? They can have an investigation ongoing, and share the cause for said investigation. In the statement they explicitly establish that this doesn't imply guilt of the account owner.

Why are you so triggered by this clarification?

Re: Google is investigating the actions of another top AI ethicist

#109

Post by visarga »

gillesjacobs wrote: post #100
visarga wrote: post #67

Earlier quoted context omitted.

It was clearly a mistake. They don't act in a constructive way. Let's take Timnit's paper for example. She found out that Google had been using a biased language model in search. The bias was like - 'male' association with 'doctor' as opposed to 'woman' association to 'nurse'. But she didn't show how this is harming anyone in a concrete way. Just theoretical. And then she offered no solution, just blaming the work of…

While I am not a fan of Gebru's handling of her own firing, this is a mischaracterization of her work. She routinely offered suggested methods, experiments, and even new datasets [1] that fixed what she saw as wrong. She does have a fair bit of practical ML experience as shown in the computer vision papers in her publication list. 1. Buolamwini, Joy, and Timnit Gebru. "Gender shades: Intersectional accuracy dispariti…

Of course I was aware of her Gender Shades paper. It's a small bias evaluation benchmark dataset.

> We developed the Pilot Parliaments Benchmark (PPB) to achieve better intersectional representation on the basis of genderand skin type. PPB consists of 1270 individuals from three African countries (Rwanda, Senegal, South Africa) and three European countries (Ice-land, Finland, Sweden) selected for gender parity in the national parliaments.

A dataset of 1270 images is hardly a breakthrough, the kind I expect to see in a small university project. But it doesn't lead to better models because it's not nearly large enough to train on. What it can do is rate existing models. Basically - useful to critique, not to improve.

A small nitpick: why just two races in a de-biasing dataset? Where are the Asians?

Re: Google is investigating the actions of another top AI ethicist

#110

Post by visarga »

jsinai wrote: post #101
visarga wrote: post #67

Earlier quoted context omitted.

It was clearly a mistake. They don't act in a constructive way. Let's take Timnit's paper for example. She found out that Google had been using a biased language model in search. The bias was like - 'male' association with 'doctor' as opposed to 'woman' association to 'nurse'. But she didn't show how this is harming anyone in a concrete way. Just theoretical. And then she offered no solution, just blaming the work of…

> But she didn’t show how this is harming anyone in a concrete way. To roll with the example “man” ~ “doctor”, “woman” ~ “nurse”, the harm is having a giant and widely used search engine reinforce baseless gender biases, ie that there is no underlying reason why women should be nurses and men doctors. What is the harm you may ask? The harm may be subtle, eg being surprised when you find out your next doctor is a woma…

This kind of hypothetical effects should be documented in concrete cases by a good ethical scientist instead of just described from imagination. For example, when someone searches for a doctor, let's say a male comes up first - who stops at the first Google result? They would probably need to go deep and read about the doctor's experience and find patient reviews.
Post reply on HN