For people who are familiar with this, I have a question. How does research in an industrial lab vs. a university work, insofar as independence is concerned? In a university, aside from well-defined ethics guidelines for conducting research (e.g., involving animal or human subjects), staff are generally free to publish the findings of their research at any venue they see fit. Many are tenured professors, so they also…
MSR has a very free research culture. People there are technically fireable by normal means, but Microsoft chooses not to terminate employment for "outside-of-work" engagements, or for the subject of their research. Some very critical papers of Microsoft's AI approach were published under the MSR header. Not all of the findings were put into place, and Aether has the unenviable job of balancing ethical behavior with…
Margaret Mitchell fired from Google
101–110 of 864 posts
Re: Margaret Mitchell fired from Google
#102- 135,000 Google employees - 2,000 signed a petition to support Timnit [1] - 2 employees resigned to support Timnit [2] When it comes to voting with your feet, no one actually gives a shit. [1] https://www.cnbc.com/2020/12/04/thousands-petition-google-fo... [2] https://www.cnn.com/2021/02/04/tech/google-employees-quit/in...
Re: Margaret Mitchell fired from Google
#103It's pretty important to get this story straight: She moved files when she wasn't supposed to. This seems like a pretty standard firing. People are already claiming on twitter that Google is assaulting ethical AI, and I just don't see it. Google did not immediately comment on those claims. Its statement said of Mitchell: “We confirmed that there were multiple violations of our code of conduct, as well as of our secur…
If that's true, that's not very ethical is it?
Re: Margaret Mitchell fired from Google
#104Earlier quoted context omitted.
Gebru wasn't fired because she "cast Google in a bad light", she gave an ultimatum to her boss threatening to resign if they don't disclose names of all reviewers who criticized her work AND blasted group emails calling for sabotage https://docs.google.com/document/d/1f2kYWDXwhzYnq8ebVtuk9CqQ... https://www.reddit.com/r/MachineLearning/comments/k77sxz/d_t...
Her research does cast Google in a bad light. She writes about how language models (like Google's BERT which influences somewhere between 10% and 100% of searches) are environmentally unfriendly and encode hegemonic language making them subtly racist and sexist. Her solutions are to use smaller language models and more curated data sets, which are probably not palatable for Google. Dr. Gebru was told the paper should…
Re: Margaret Mitchell fired from Google
#105- 135,000 Google employees - 2,000 signed a petition to support Timnit [1] - 2 employees resigned to support Timnit [2] When it comes to voting with your feet, no one actually gives a shit. [1] https://www.cnbc.com/2020/12/04/thousands-petition-google-fo... [2] https://www.cnn.com/2021/02/04/tech/google-employees-quit/in...
Re: Margaret Mitchell fired from Google
#106Earlier quoted context omitted.
Releasing even a single internal document without permission is cause for firing, and possible legal action. I can't even fathom this being up for debate.
Not necessarily. Google's US-based employees would be protected by Sarbanes-Oxley Act if they did it to report on fraud. But generally, yes. Unless you have a truly extraordinary reason you can get easily fired in most companies.
Re: Margaret Mitchell fired from Google
#107Re: Margaret Mitchell fired from Google
#108Earlier quoted context omitted.
> Had someone on my team decided to write a similar doc like the one Margaret published, I'd have fired them on the spot I sometimes don't agree with decisions my (present and past) employers have made. But I'd never go to the media or public forums with those disagreements. If I thought it was so important that I had to do that, I'd resign first. I strenuously avoid making any public comment about my current employe…
sorry about your backbone
Instead, if you'd please make your substantive points thoughtfully and without taking the thread further into flamewar hell, we'd be grateful.
Re: Margaret Mitchell fired from Google
#109Google response (from Jan): "Our security systems automatically lock an employee’s corporate account when they detect that the account is at risk of compromise due to credential problems or when an automated rule involving the handling of sensitive data has been triggered. In this instance, yesterday our systems detected that an account had exfiltrated thousands of files and shared them with multiple external account…
That seems like a pretty serious accusation. Looks like relations between management and employees has seriously deteriorated.
That said, anything that makes Google a less effective organization I support.
Funny how they co-announced monetary incentives around "diversity and inclusion", I expect nothing less from a giant evil ad company. A bunch of handwavy idpol stuff with the giant elephants in the room seems to be the status quo.
Re: Margaret Mitchell fired from Google
#110Now there is a line between public and private space, but many people use Twitter as a professional network. And Dr. Mitchell includes her Google credentials in her bio.
So given her substantial, critical, public views of her employer on a quasi-professional network - what did she expect?