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

Margaret Mitchell fired from Google

axios.com

Post reply on HN
Print view View original

791–800 of 864 posts

Re: Margaret Mitchell fired from Google

#791

Post by jariel »

5600k wrote: post #430
belltaco wrote: post #409

Earlier quoted context omitted.

According to an Axios scoop from Jan, she was running a script to find emails in her Inbox related to Gebru and sending thousands of files "to multiple external parties". https://i.imgur.com/fDV9kKV.png

Would anyone suspect that an AI set her up? One could be fully capable of that. This whole thing may be ironic- in more than one sense. I’m kidding of course... or am I?

I don't see how anyone 'set up' anyone.

Paradoxically, AI may have been part of the tech used to flag her, but it's all mundane at this point.

Kind of implying that AI is just 'tech' and there's no material reason to separate AI from everything else from an ethical perspective.

Google's main Search Engine should be the biggest point of controversy if there is one - the things they chose to filter, or not, or highlight, their ranking, etc. - a lot of that is algorithmic and derived from human input and it has massive impact. But there's no cool moniker like 'AI' with Hollywood movies about the tech to get people paying attention.

Re: Margaret Mitchell fired from Google

#792

Post by jariel »

randycupertino wrote: post #651
totalZero wrote: post #551

Earlier quoted context omitted.

Some of these "AI ethics researchers" seem like wingnuts, and their profession appears to serve more of a PR purpose than a business purpose. What am I missing? Why are they so essential? Are these people the canaries in the coal mine, or do they simply exist because of paranoia and political correctness?

So, I work in medical research and was headhunted by (but turned down) Juul to work on sponsored trials and research related to smoking cessation and health improvement in patients who are switching to pods. Per the article, "Google has recruited top scientists with promises of research freedom, but the limits are tested as researchers increasingly write about the negative effects of technology and offer unflattering…

Different companies, different industries, different rules.

Researchers can publish anything they like as long as it's uncontroversial. Google sponsors tons of work for which there is no oversight.

But if you're going to directly make the company look bad, then yes, it's delusional.

It's one thing to protect whistle blowers, it's another thing to make bad public statements. The previously fired researcher wanted to demonstrate how training AI was energy intensive and that was a problem, though in the grand scheme it may not be i.e. it's not as though model training is a primary source of energy draw. Google wanted to review the work and asked her to include more recent information about how Google had actually improved energy consumption process given the old data set from the research.

That's perfectly reasonable. Corporate HQ asks the researcher to include more recent data which doesn't make the company look so bad? On a scale of 1-10 in terms of interference that's a 1 or 2. I suggest the vast majority of researchers would be fine with that, and consider it ethical.

Juul probably hires scientists more along the lines of cigarette companies wanting to provide science to show that cigarette smoking is not harmful. That's another thing altogether.

Re: Margaret Mitchell fired from Google

#793

Post by Cacti »

pseudalopex wrote: post #774
Cacti wrote: post #759

Earlier quoted context omitted.

Everyone is assuming it’s email, but you don’t need “a script” to search your own email, and you don’t need to risk getting fired sending emails to your lawyer, you just print them out. Or wait to get fired and get them during discovery. Google isn’t going to go into conspiracy mode and risk hundreds of millions over a line manager getting fired, that’s absurd. It wouldn’t surprise me if this “script” was trying to p…

Saying documents could mean emails doesn't assume it's email. Axios said the unnamed source said Mitchell looked through "her messages" though. And what other relevant files would there be thousands of? Google Apps Script can do things Gmail search can't. Like regex. Lawyers I know say don't rely on discovery if you can help it. You think Google doesn't monitor printing?

I mean, generally speaking you can’t get fired for printing your own emails.

Re: Margaret Mitchell fired from Google

#794

Post by monocasa »

JamesBarney wrote: post #777
monocasa wrote: post #639

Earlier quoted context omitted.

> I'm not trying to prove she "voluntarily quit" vs was "discharged" via California law. I'm arguing that Google used resigned in a way that wasn't a lie, and referencing California law to show how it uses the word resignation, and is therefore a reasonable use of the word. You can't first claim that you're not using California law in your argument, then immediately cliam that your arguement is based in California la…

The point is for someone to be lying their use of the word has to be inconsistent with how any reasonable person could use the word. My argument is that if you use the word in the way California law uses it your not lying. But also if you don't use it in the way California law uses it but the way U.S. law uses your not lying. Or if you use it in a manner that's inconsistent with California law and U.S. law but is how…

> But also if you don't use it in the way California law uses it but the way U.S. law uses your not lying.

You've given no citations that "US law" is somehow different than California law here. I've stuck to the specific citations directly applicable to their situation, but the idea that if the employer is the one who specifies the end date of employment, it's a firing not a resignation is pretty universal across the western world.

And yes, using "everyday language" to make public statements about a legal situation that if you used the legal definition you'd be making the opposite statement is very much a lie.

Additionally, even the "everyday language" argument has you making a distinction between "voluntary quit" and "voluntarily leave a job or other position". The crux is that if the employer specifies the end date, it's not voluntary on the part of the employee. Both under legal definitions and common parlance.

Re: Margaret Mitchell fired from Google

#795

Post by jgalt212 »

Honestly, who's the bone head who set up this department in first place? It's like letting the cock sleep in the henhouse. Either this department is one big charade, or it tries to do a good job and is in constant and acrimonious conflict with rest of the org.

It's like a bank or HF having a compliance department who is not only responsible for ensuring rules (it doesn't entirely understand) are followed, and making up new rules at the same time. In effect, a dual legislative and judicial branch.

Re: Margaret Mitchell fired from Google

#796

Post by bryan0 »

notsureaboutpg wrote: post #423
bryan0 wrote: post #419

Earlier quoted context omitted.

“Cut and dry”? Given how google fired a previous AI ethicist in an entirely misleading way (“We accept and respect her decision to resign from Google”), it would be naive to take google at their word here.

Of all the things to argue about how that incident went down, this is the worst. She sent an email with a list of demands saying if they weren't met, she'd work on an exit date with the company. The company couldn't meet those demands. By her own words she would be exiting the company, they worked out that exit date with her in that email: right now

Yes, I realize I could have picked from several other examples which demonstrated Google acting in bad faith, but I felt that quote (which I put in parens and was not the main point of my argument) was the most memorable from the whole sad affair.

Jeff Dean just yesterday officially apologized for this situation so clearly he feels he did something wrong, although he doesn’t admit to anything specific. It wouldn’t surprise me if this quote from his email (and the corresponding behavior) would be on the list though.

Re: Margaret Mitchell fired from Google

#797

Post by prepend »

geofft wrote: post #723
valdiorn wrote: post #656

Earlier quoted context omitted.

She's also not publishing this as just herself, she works for Google and is publishing on their behalf. Of course they have veto power over what she publishes. You take issue with that; quit. Which, I believe she did (she didn't get fired AFAIK)

Yes, there exist review processes by peers, and yes, your institution can choose to exercise veto power. The scientific process of "peer review" is a very specific thing. (My code is not "peer reviewed" in the scientific sense simply because a teammate, a peer, reviewed my pull request.) So saying the paper was not good enough for peer review is misleading. And yes, Google can veto any paper. That's not the issue. Th…

My organization includes a peer review process before it is submitted to journals. It makes sense because how would an organization be able to do this without peers.

There are multiple types of peer review, not just from formal journals. Your statement of a narrowly scoped peer review being the only kind is incorrect.

I’m not talking about Google specifically, but I imagine any institution creating scientific knowledge will have multiple phases of peer review before the literature.

Separately your logic of “no paper coming out of Google can be trusted” is faulty. It’s not that it can’t be trusted at all, it just can’t be trusted for some things. Peer review outside Google should identify if Spanner has bugs or whatnot. I think it’s safe to say that work of Google should not be estimated to only exist based on what’s published.

So while Google published things should be true since they are reviewed outside of Google, we’ll never know the things not published.

Also this is sort of the point of science right, you shouldn’t blindly trust things based on institution, but need to review critically based on your own experience and expertise.

Re: Margaret Mitchell fired from Google

#798

Post by michaelcampbell »

im3w1l wrote: post #577
carapace wrote: post #481

Earlier quoted context omitted.

Instead of getting rude, you could have just searched "one finger points out, three point in" on DDG: https://duckduckgo.com/?t=ffcm&q=one+finger+points+out%2C+th... First result: > 'When you point a finger at someone, remember there are three pointing back' ...

What makes it hard for me to read is mainly the comma-placement, which is quite different from the verbal pauses you'd make when reading it, and break up rather than separate the logical parts. I found myself asking "what does it mean to point in rule ?" Italics or quotes would have helped.

I think it was more the lack of any sort of container; does:

...the "one finger points out, three point in" rule, generally applies...

parse any better?

Re: Margaret Mitchell fired from Google

#799

Post by prepend »

wpietri wrote: post #717
prepend wrote: post #441

Earlier quoted context omitted.

No one in this scenario reported an ethics breach right? So how would it apply? There are whistleblower laws now but they apply to illegal actions.

The heart of the matter is what the ethics experts see as an attempt to suppress criticism of ethical problems with Google's technology. So yes, it's very much an ethics breach. Whistleblowing need not be only about crime. It can just be about harm. [1] Indeed, with new technologies, regulation often has yet to be written, so whistleblowing can only be about harm. [1] https://blogs.scientificamerican.com/observations…

I think that’s the disagreement.

Dean said in his letter that the paper didn’t address relevant literature related to the subject that the authors weren’t aware of. To me it doesn’t seem that the reason is that it besmirched Google, but that it wasn’t complete.

Also the article wasn’t an ethics breach type of article and internal so I’m not sure what the legal protection would be. I think if the employees had some evidence of illegal activity then they would be protected.

If this paper would be considered an ethics breach then I don’t think there should ever be relevant protections.

Re: Margaret Mitchell fired from Google

#800

Post by michaelcampbell »

renewiltord wrote: post #279
madamelic wrote: post #39

Earlier quoted context omitted.

If you fire your "Ethical AI" team lead for disagreeing with you... maybe you are only interested in the appearance of being ethical. Just a thought.

There's room for discussion, certainly. If I hired an accountant to keep me on the straight and narrow and I expensed a dinner I'm not supposed to, I wouldn't want him to go running to Twitter and be like "Hey, everyone! Rene expensed a dinner with this guy abroad! That's a violation of the FCPA! He's a violator, guys! Heeeeeee's a violator!" Like, if he did that, I'd be pretty incensed honestly. I'd expect him to te…

> There's room for discussion, certainly. If I hired an accountant to keep me on the straight and narrow and I expensed a dinner I'm not supposed to, I wouldn't want him to go running to Twitter

Or emailing your expense report to who knows whom.

Post reply on HN