17 ms·
Google told its scientists to ‘strike a positive tone’ in AI research
- xiphias2 6y agoWe already know that ,,self regulation'' doesn't work. It made sense to me that when I was working at Google I didn't talk bad about the company that I was getting my paycheck from. Alphabet is still a great company, but that doesn't mean that it doesn't need checks and balances from outside.
- dls2016 6y ago> I didn't talk bad about the company that I was getting my paycheck from Have you ever been in a union workplace?
- throw_m239339 6y ago> Have you ever been in a union workplace? why would he want to be in a union when he is a top earner compared to great majority of people in the same country? and aside from a relative freedom of speech he works in very favourable conditions with a lot of perks and a generous severance package if he gets fired? working at google is the very definition of privilege.
- ClumsyPilot 6y agoMaybe then he would not have to tow every pr line in fear of being dismissed
- ThrowawayR2 6y agoHeaven help him, though, if he says anything critical of the union. Tire slashing and other forms of intimidation routinely made newspaper headlines in UAW territory in the era before the Big 3 automakers collapsed. At least dismissal doesn't involve people following you to your home. It isn't businesses that make US unions look bad, they did it to themselves.
- dls2016 6y agoWhat other balance to corporate power do workers have?
- marcus_holmes 6y agoUnions are for more than just getting higher pay. If Google (perish the thought) decided to pull some shady reviewing stunts and make your work life so intolerable that you felt your only option was to force the issue and threaten to resign - being part of a union could have helped with that. HR is a bunch of people who represent the company's interests in negotiations and transactions. Unions are the same thing for employees. We've been told, repeatedly, that unions are socialist organisations that want to destroy capitalisim and bankrupt employers. This isn't true.
- whimsicalism 6y ago> union when he is a top earner compared to great majority of people in the same country Why does that make creating a union/engineer council less desirable? It seems good to have a democratically elected representative of Google engineers who can talk to management, even if those engineers don't feel like negotiating on wage stuff, they still have a real avenue to air concerns.
- dls2016 6y ago> why would he want to be in a union when he is a top earner compared to great majority of people in the same country? Didn't Google get busted fixing wages along with other tech companies? https://en.wikipedia.org/wiki/High-Tech_Employee_Antitrust_Litigation https://en.wikipedia.org/wiki/High-Tech_Employee_Antitrust_L...
- xiphias2 6y agoNo, but my work life was amazing. There are important things to address at Google, like not having enough sitting toilets for men (there's 3 times as much for women even though there are much less women in the office where I was working) that an union may help with, but that's different from the product PR.
- mosselman 6y agoForcing men to poop standing up is probably some productivity hack.
- purple-again 6y agoMy first inclination was just to move on but you didn’t post on a throwaway so I will take the bait. Why do you think the toilet issue you described is a problem that needs to be fixed? My experience in every corner of life in America to date has been separated bathrooms by sex and the men’s restroom containing far fewer sitting toilets. We all know this is because women sit for both natural duties and men, generally in the US at least, only sit for one. This is so basic I assume it’s not what you are talking about but please correct me if not so we can engage on that front. What I assume you mean is Google is 90% male and 10% female (no idea what it really is) and yet the bathrooms are 50% male and 50% female. You specifically mention sitting toilets which is what I am having trouble understanding the relevance of. Are men regularly forced to stand in line instead of being able to walk right in and go? If so that sounds like a facilities growth issue and not a sex imbalance issue. You could propose turning some female restrooms into male but generally restrooms are placed far enough apart from one another that there would have to be 0 females in the area serviced by those restrooms for that to work. The presence of even one would require facilities assigned to females or mixed gender restrooms (a different argument I fully support). That leaves your only other option as I see it for the female restrooms to be smaller than the male restrooms which is just ridiculous. No one would build the building with that hopefully temporary imbalance in mind and you can’t just repurpose half a bathroom as long as the sexes are being separated.
- extropy 6y agoI will bait too, as someone who has stood in line for male toilets in a Google office. In the office I was in there where generally 1 sitting and 2 standing seats per male toilet. The standing one are quick in / quick out and almost always free. But good luck finding vacant sitting spot after lunch. Edit: female restrooms generally had 3 sitting spots, so same size.
- sgregnt 6y agoCan you elaborate? Who are "we", does this refer to some scientific community studying these things?
- cmiles74 6y agoIn my opinion, this is different: Google is hiring research scientists and telling them that Google would like to pay for their research. When hired, Googled told these people that they would be free to conduct research in their field without this kind of censorship. Now they are being censored, this is a new policy. In my opinion, this is very different from a typical employee "talking bad" about their employer. "Google’s new review procedure asks that researchers consult with legal, policy and public relations teams before pursuing topics such as face and sentiment analysis and categorizations of race, gender or political affiliation, according to internal webpages explaining the policy."
- jariel 6y agoAny researcher who thought they could publish research about 'race and gender' and not have it reviewed by a corporate sponsor is incredibly naive. It's an extensively sensitive subject and particularly in a field with such vague and 'subject to interpretation' results, the sponsor corp. is going to want to have a look. Considerably more so if the research is about the institution itself. There was a specific case recently where the researcher criticized Google for something, but Google asked the researcher to include truthful information about the changes the company has already made to the system in question wherein it's been improved etc. This is a perfectly fair request by the sponsor, I think. It's also naive for us to consider that what people chose to research, how they approach it, the results they are looking for isn't somehow altered by the researchers own biases. In fact, it's hugely arrogant for researchers to believe that they have some special access to the truth in this regard.
- cmiles74 6y agoPerhaps researchers who accepted these positions with Google were naive. That is certainly possible but it can't be argued: Google hired these people under the pretense of letting them do their research free of censorship and, now that they work for Google, this censorship is being imposed. I'm not sure what the rest of this comment is alluding to... The posted article from Reuters doesn't mention race and gender, the article implies these researchers focus on AI and ML. In regards to the "specific case", I would have to guess at what that case might be. Were I to guess, it seems like I'd be guessing wrong as this assertion seems to lack grounding in fact. Naive or not, peer review is in place and part of it's purpose is to keep the biases of researchers in place. History has proven that this is far from infallible, but it's the process we have today and the process Google agreed to when it hired these researchers.
- Bakary 6y ago>Alphabet is still a great company To those it pays, yes...
- curiousllama 6y agoIf the outcomes of research don’t align to Google’s business strategy, they’re going to get rid of the research/researchers. Previously, ethics was good PR; now, it’s a risk in antitrust cases. This will continue to ramp up.
- oblio 6y agoIt will also continue to ramp up even more as growth dries up. The internet can only get so big, especially on a quarterly basis, and these companies are already very big. They're public and pressure to grow is huge. It's going to be a rough ride throughout the next decade.
- microtherion 6y agoYou may well be right that this is Google’s current thinking, but it doesn’t look like a winning strategy to me. Once more, they are on the defensive trying to offer flailing justifications for an opportunistic initial move. Even if you don’t care about the ethics, or are on their side, I think you’d have to question their execution.
- curiousllama 6y agoOh, I don't think these are post hoc justifications. Much more likely Gebru was intentionally forced out - I wouldn't be surprised if Google leadership basically said "she's a loose cannon - move her out in the next 6 months" and there was just no good way.
- jariel 6y agoNo. More like, researchers can do as they please, but if they're going to attack Google or be negative about something related to Google, then there's going to be a review and words will have be chosen carefully. There's a big difference.
- amitport 6y agotl;dr;: + "Google’s new review procedure asks that researchers consult with legal, policy and public relations teams before pursuing topics such as face and sentiment analysis and categorizations of race, gender or political affiliation, according to internal webpages explaining the policy." + "A senior Google manager reviewing a study on content recommendation technology shortly before publication this summer told authors to 'take great care to strike a positive tone,' according to internal correspondence read to Reuters." + surprisingly, there's not much else in the article beyond those two sentences Edit: as commenters noted, there is more in the article. I apologize for not being clear that the last bullet is just MY OPINION. That is, for me the rest is not as meaningful as some consider it. I guess, if you care deeply on the subject, you'll want to read the article and decide for yourself.
- daniel957 6y agoWhy does your last bullet point make me feel like we read two different articles?
- mlthoughts2018 6y agofwiw I agree with the parent comment. This article, especially the ludicrous attempt to tie in the Gebru stuff, is super weak and absent of any details or anything substantial. I don’t see anything in the article that suggests Google suppresses unfavorable research or censors authors whatsoever. Instead it sounds like a lot of whining and complaining from exceedingly privileged researchers who want some version of tenure to apply within for-profit companies and seem too immature to accept that legal and compliance functions within a large corporation very differently than in academia.
- daniel957 6y agoWhy did you find the mention of Timnit Gebru ludicrous? The issues are closely related. In fact, I was _expecting_ the article to mention it. Google tells Researcher 1 to re-word things in X way for Y reason. Google tells Timnit Gebru to re-word things in X way for Y reason. The X and Y are the same. Based on the rest of your response, I believe you are having trouble reading the article in an objective manner. You say that it sounds like whining to you. That’s because you already think it is. You came into the article with this viewpoint about academics in corporations already established. This is why I also found your comment hilarious. You already think Timnit is a whiny researcher. You say this article is about whiny researchers. Yet it’s _ludicrous_ to mention Timnit in this article? Also, Jeff Dean lied in his email. Maybe that’s why you’re confused about the Timnit situation. I’m assuming you don’t work at Google.
- axegon_ 6y agoHah, reminds me of this literature teacher I had in school. She asked us to write a report on the positive aspects of Hamlet. I remember sitting at home and thinking "Well... It could have been worse - more people could have died, that's positive thinking, right?"
- mistersquid 6y ago> She asked us to write a report on the positive aspects of Hamlet. Spoiler Warning! Rosencrantz and Guildenstern get what is probably their just desserts. If that’s too uncertain given they may not have known the fate the King had in store for Hamlet, then Fortinbras. Fortinbras is a pretty standup guy through it all and his closing lines provide the possibility of a better future.
- rabbut 6y agoIn Terminator Genesis, the AI was put in a good light also- on billboards and such. It didn’t work in that movie, though. The AI still ended up being hostile even when the public embraced it. Why are we xenophobic about AI self-awareness? idk
- throw_m239339 6y agoGoogle is a for profit corporation, not a college campus, news at 11.
- dmingod666 6y agoPeople will act like this is up for debate.
- square_usual 6y agoEverybody knows that. You can think it's a bad thing for a private company to do bad things, even if they're free to. Stating a fact isn't an argument.
- izacus 6y agoI think the comment flies more in the wierd way Google is treated by the media when other corpos aren't. E.g. noone is batting an eye when Apple chooses cherry-picked based datapoints for their performance comparisons or when Stanfords research department spins a paper to get them more funding. For Google we keep getting this streams of "Google is acting like any other corporation, OH GODS". It's... wierd.
- SamoyedFurFluff 6y agoI dont know if a case where Apple hired ethicists and then fired them for their research? Admittedly apple also doses all their employees with NDAs up the wazoo so we simply can’t know these things.
- unishark 6y agoAll I'm asking for is the freedom of tenure combined with the pay and equity of industry.
- bonoboTP 6y agoThe top dogs have both, with multiple affiliations.
- drfuchs 6y agoConsider the historical parallel to the tobacco industry’s research departments. And asbestos. And lead. And talc. And etc. The researchers found that the products were harmful, but their results were squelched. “Do no harm,” indeed.
- ruslanbogan 6y agoThe difference is that AI research teams are full of philosophy graduates who think if statements are racist
- oblio 6y ago> The difference is that AI research teams are full of philosophy graduates who think if statements are racist. I'm not sure I understand. Are they worse or better?
- raverbashing 6y agoAlgorithmic biases and their application to human systems are a very interesting, and important topic. And yes, they can be unfair to already disadvantaged people But I think the discussion too easily devolves into the craziest humanities self-pleasing circles. Then they are surprised they get push back? Of course Google also share their side of the blame when they think just creating an "ethics commission" with a rubber stamp solves their issue
- piokoch 6y agoThat's an actually good question. AI is supposed to be fed with real-live data, after all it has to be something usable in the real-world. Now, you feed AI with crime data, looking at statistics such AI will suspect people having black color of the skin to be criminals more often than people with white skin. Is that ok? Is it ok to say that if some person happens to have black skin there is a bigger chance that it is criminal? In statistical terms the answer is yes, but if we deal with an actual person, we should treat it as a unique person and the matter of the skin color should not matter. Before II World War there were scientific studies in Germany that showed that in some big cities districts there are more rats. Accidentally or not, it turned out that those are districts with large Jewish population. So Nazis took those statistical results to prove that Jews are dirty, spread diseases, etc. Was that science, even though it was peer reviewed and published? Rather not (in the first place they didn't even bother to check if people living there were rich, poor, etc.). History knows many "scientific researches" by the ones like Eugen Fisher, dr. Mengele, that were performed in a scientific way, but they were lacking something very important and non-trivial: ethics. Now the problem is that ethics needs to have some source, it does not grow on a tree. For ages Christianity was a source of ethics in Europe (less in US, that's why you could have slavery in modern times there), currently Christianity is almost gone in Europe, the question is were to seek the source of ethics.
- nisuni 6y agoAnd that’s why I was happier when basic research was mostly being carried out in public institutions.
- _Nat_ 6y agoPublic institutions tend to observe the same policies, even when there's no direct commercial interest. For example, if you join a Math department and start publishing papers that cast the department's research interests in a negative light, that's often a good way to get fired. I mean, seriously, go ask just about any professor of any topic at any university if their field is undervalued or overvalued. They'll almost universally tell you that their field is undervalued, underappreciated, and more important than people think. They deserve more funding, and you should definitely consider majoring in their field.
- kovac 6y agoGeneral proposition may be true, but how can a math publication place the department interests in a negative light? Like "Doesn't matter how hard we try, there are always undecidable facts in our own system"? :p I still have faith in the research done by hard science research teams.
- _Nat_ 6y ago> how can a math publication place the department interests in a negative light? With Math, the sensitive topic is utility. Math departments would tend to be upset by someone pushing papers that, say, call attention to the lack of social benefit relative to other fields of research that taxpayers could be funding instead. I guess a researcher could also spin a narrative about mathematicians' contributions to cryptography, examining the negative consequences enabled by stuff like Bitcoin. --- > I still have faith in the research done by hard science research teams. That gets kinda complicated. I mean, hard-science teams often rely on their expertise, as it's their unique source of value: if the area that they excel in is shown to be sub-optimal, admitting it would mean losing their current career path (and often taking a significant tumble down-the-ladder after transitioning). So hard-science teams are often largely academically honest in what they do publish, though they're often biased toward casting what they do in a positive light. For a common example, have you ever met an older computer tech who insists on using a legacy, obsoleted technology because it's what they know best? A lot of academics basically do the same thing.
- _cs2017_ 6y agoIMHO, a scientist capable of striking any tone in their work, positive or negative, is doing marketing / PR work rather than science. There's nothing wrong with that. Just like any other member of society, a scientist may want to contribute to a (for-profit or non-profit) organization of their choice, be it for material or spiritual rewards. However, I think it's important not to get science and marketing confused. So, I think a more precise title would be "Google told its scientists to strike a positive tone when they use academic papers as a marketing tool".
- FemmeAndroid 6y agoAny scientist writing prose will necessarily strike a tone. Unless they’re just publishing raw data, the way you use that data in your prose will lead to a different tone. I have read a good deal of historical and modern scientific papers and they all had different tones.
- chaboud 6y agoWhen I was young and (more) idealistic, my department chair pulled me aside after I’d been dogmatic about something and said “everything is political... The sooner you realize that, the better off you’ll be.” It took 10 years for that to really sink in, but I knew that every paper has an angle well before that. There’s a position of inevitability about AI around Google, and, to an extent, I agree. Unlike tobacco, alcohol, firearms, cars, fuel, AI is applied mathematics. It’s virtually impossible to restrict functionally, requiring macro behavioral constraints on businesses and governments that are very difficult to enforce if you want to regulate it. Furthermore, ceding the technological lead and removing the business incentive likely has defense (read: offense) implications on the global stage. Some things are going to be hard to explain. Some skin tones, hair styles, accents, pitches, timbres, writing styles, etc. may have more/less separable entropy, and that’s going to inevitably make models seem racist/sexist for some applications. However, we’re pretty far from those edges. Instead, our model biases have likely been driven by our datasets, drawn from our coworkers, family, and friends. In this industry, that’s currently a recipe for over sampling in some demographics and under sampling in others.
- 6y ago
- tanilama 6y agoIt is at will employment, the company makes it clear what you have done there, they want to have a say on it. If you want independence, goes to academia instead (though it is not easy to obtain it either)
- glitcher 6y agoIt is at will research funded by google, the researcher makes it clear what their findings were, they want the original research paper to have its say based on results not spin. If google wants specific outcomes from research, go to the marketing department instead of funding real research.
- spoopyskelly 6y agoWhy would you want to shit on the company you work for? Just go virtue signal elsewhere.
- xpe 6y agoYou are offering a false choice / false dilemma: https://en.wikipedia.org/wiki/False_dilemma https://en.wikipedia.org/wiki/False_dilemma Many of the comments on this thread are expressing concern and (more or less) asking 'why does it work like this?' and (in some cases) 'it would be better if it didn't work like this."
- DrNuke 6y ago2021 will be an interesting year and the feedback I got after my latest booklet / lit review aimed at Silicon Valley problems is just explosive: tightening regulation, legal actions, anti-monopolies concerted effort will ramp up dramatically at worldwide level.
- op03 6y agoPoor Google managers. Nothing they do, say and think is private anymore. How sad :)
- 6gvONxR4sf7o 6y agoIf only there was a strong privacy culture there.
- LoSboccacc 6y ago> internal correspondence read to Reuters so, all of this opinion piece is based on hearsay?
- Closi 6y agoHow do you think journalists get stories like this if not from insider sources?
- LoSboccacc 6y agojournalist should at the very least receive, see and validate documents veridicity. here it's "internal correspondence read to Reuters." from a highly politicized topic with a lot of disinformation being thrown around from both side, it's hardly sufficient. but apparently you all happy enough to live in a post factual society where journalist have no longer to validate sources' claims, so be it
- Closi 6y agoLol, this is how journalism has been done for decades - There is nothing 'post factual' about it.
- LoSboccacc 6y agoyeah no, imagine watergate without the address book, the tapes and the bank records.
- hajderr 6y agoIdea: AI bot for detecting positive tones in Google's papers.
- unishark 6y agoHow does it make you money? My money-making idea from this is AI to detect sentences or fragments from your emails that could be taken out of context and made into embarrassing headlines if the email gets leaked.
- tim333 6y agoDon't all corporations want people working for them to be upbeat rather than calling cars death machines etc?
- mlthoughts2018 6y agoThe “sensitive topics” extra review seems completely fair and obviously necessary. Anyone acting like that amounts to censorship in a private company is totally off their rocker. This article seems really weak, and of course there’s the obligatory attempt to tie it in with Timnit Gebru’s resignation and her very inappropriate behavior, even though for all of Google’s flaws, Google clearly and obviously did the right thing both to disapprove Gebru’s weak paper and accept her resignation based on her ultimatum.
- ClumsyPilot 6y ago"Gebru’s weak paper" Could you please back up your claim as to the paper's quality - do yoy have credentials and background that enable you to cast this judgement, or is this pedestrian?
- free_rms 6y agoI don't need credentials to read a paper which muses about current events. You want to outsource your thinking to the credentialed? Go ahead but the rest of us don't have to. The paper is out there and it's quite easy to read, because it's not saying a whole lot.
- bonoboTP 6y agoIt has leaked. Activism masquerading as science. Yes, people also call critical race theory and intersectional theory academic disciplines. They hold conferences, have titles and credentials, are professors, etc. So yes the paper is high quality for a certain niche. It's not a technical paper for not a technical conference. It's a "raising awareness" paper on identity politics beating the same drum, without much novel insight or analysis or new data. But yes, it has 120 or 200 citations, a fact that was proudly highlighted by authors on Twitter to underline the scholarly quality. Yes, gender seminars and whiteness workshops are also properly credentialed. Theology is also a discipline but perhaps not the same as sciences. Quality is indeed relative.
- 6gvONxR4sf7o 6y ago
- frankfrankfrank 6y agoI deal with this on a regular basis in the research world, there is what is a really toxic and dangerous bias towards constructing the positive, i.e., making things seem, appear, and even fabricate outcomes that will be affirming of desired outcomes or other group biases. It is not research, let alone science when you are striking a positive tone, it’s fraud.
- curiousllama 6y agoThis is a really important point. Executive-land is all about image and framing, so I can totally see how execs think it's not a big deal to "just make the case" or frame things less negatively. But while that's ok in a slide deck, it's _not_ ok in a research paper. Researchers aren't (shouldn't be) accountable to business outcomes, but rather the truth of what they argue. It's tough to see this play out real time.
- tomcam 6y agoGrant-land is very very often about image and framing. How often are grant administrators interested in pure research and let the chips fall where they may vs getting results that yield happy articles, then more grants?
- ludamad 6y agoAgreed, the bias towards culling negative results is not unique to Google, although the mechanism may be
- silexia 6y agoThere is a witch hunt against Google that is politically motivated happening now. If Google is funding its own research, it should be able to look at whatever it wants. Why do we hold Google to a standard that most other companies are far worse at following? The appropriate place for independent critical research is outside of the company itself. Google is actually a pretty well intentioned entity to put out as much self critical research as it does.
- oh_sigh 6y ago> The manager added, “This doesn’t mean we should hide from the real challenges” posed by the software. If I ever plucked a single sentence out of a long post and called attention to it context-free, I could probably make whoever posted it look fairly bad. But, in the context of news articles, it becomes the headline? Was the context "Try to strike a positive tone...or else you're fired" - or "Try to strike a positive tone...you've shown that these problems are solvable"
- _Nat_ 6y agoIt's important to draw a distinction between positive tonality and censoring negative results. For example, the recent COVID-19 vaccines: they're commonly reported to be 95% effective and have some risk of side-effects. This can be reported factually, or spun in a positive or negative light. For example, a sensationalist might try to stir up interest by spinning a narrative that focuses on the vaccines being unreliable and dangerous, as proven by science that Big Pharma doesn't want you to know about! And as we've seen, sensationalists can get cult followings. However bad their ploy may be for society overall, the sensationalist themself may stand to profit. So as reprehensible as censoring legitimate scientific research would be, a directive to "strike a positive tone" sounds like it could be a very different, far more sensible thing.
- whimsicalism 6y ago> And as we've seen, sensationalists can get cult followings. We're talking about published research papers, not people posting stuff on twitter. I have not seen anyone get a "cult following" from a research paper.
- _Nat_ 6y agoIt's common for researchers to publish a journal article in a more academic tone, then use it as a basis for legitimacy in public venues. This happens at both individual and institutional levels. Examples: 1.) At an individual level, someone might publish a study on a social issue that's important to them, adhering to relevant academic standards, then establish a popular identity as the author of that study. Afterwards they can comment on blogs, in media releases, on talk shows, etc., using the perceived credibility established through academic publication to pursue their personal goals. 2.) At an institutional level, someone might publish a study on some issue relevant to an industry, then use it in political lobbying related to that industry as well as popular press to build good-will. Many academics have industrial sponsors for this reason.
- whimsicalism 6y agoWell, I'd be fine with Google putting its thumb on the scale in terms of how researchers engage in public press conferences/media releases/talk shows, etc. What we're discussing, however, is "striking a positive tone" in actual research papers.
- bite_tongue 6y agoSeems reasonable for a company to not fund research that directly goes against their bottom line.
- cmiles74 6y agoIf Google did not want to fund research that conflicted with their bottom line, they could have avoided hiring researchers interested in that kind of work. Instead they hired these people and told them they would be free of censorship. Now that these people work there and depend on Google for their paycheck, Google is instituting this new policy. During a global pandemic. Sounds like censorship to me: hire the people in the field who might publish research you disagree with, then pressure them from a position of power.
- Veedrac 6y agoYou're confusing the ideas with their presentation. Google never allowed someone to post a paper under their name saying ‘datacentres generate CO₂, that's evil and we should focus on smaller datacentres’, but that doesn't mean they didn't care about CO₂ production; clearly they did, since they're now carbon negative. Same thing here. The researchers' jobs aren't to publically deface Google, they're to help direct Google down the right path.
- cmiles74 6y agoI agree that Google may have been interested in hiring these researchers to help "direct Google down the right path". That seems reasonable to me. I'm not seeing where I confuse "the ideas with their presentation" or where anyone talked about researchers "publicly defacing Google". The Reuters article is talking about ML and AI researchers who are now (recently) finding their research subject to a new review policy, above and beyond the typical peer review that they expected.
- 8note 6y agoWhat's Google's going to care about is "are we going down the most profitable path" though. It's a bit different from the right path
- nr2x 6y agoInteresting opsec: “according to internal correspondence read to Reuters”. Seems like people are afraid of a fwd paper trail.
- tartaglia 6y agoThen, is academia a good place to go to for research instead? If not, what should researchers do?
- jgalt212 6y agoThe problem is that Google wanted to monetize AI via marketing itself to a higher share price and more panache (sell the sizzle). If AI were solely an internal tool at Google, none, or very few, of these ancillary problems would have presented themselves. Or, they could have just labelled all these efforts with the less sexy and scary moniker of "machine learning".
- known 6y agoAI should be immune to https://en.wikipedia.org/wiki/List_of_cognitive_biases https://en.wikipedia.org/wiki/List_of_cognitive_biases and it's not possible;
- jchallis 6y agoEvery specialty has taboos and learning what they are and how to avoid them is part of being successful with the job. As a physics graduate student I made an impassioned case to a high energy researcher that the money spent on high energy research was low yield and should be pushed to the biological sciences ... suffice it to say he did not take my points on their merits. Upton Sinclair’s famous “ It is difficult to get a man to understand something, when his salary depends on his not understanding it.” applies in the sciences.
- aaaxyz 6y agoReminds me of an interview a cosmologist from my university gave to the CBC around the time the higgs boson was confirmed by the CERN. When asked about real world consequences of the discovery, he answered "None."
- ativzzz 6y agoI mean Heinrich Hertz, who proved Maxwell's electromagnetic wave theory also said "It's of no use whatsoever[...] this is just an experiment that proves Maestro Maxwell was right—we just have these mysterious electromagnetic waves that we cannot see with the naked eye. But they are there." [1] How wrong he was. [1] https://en.wikipedia.org/wiki/Heinrich_Hertz https://en.wikipedia.org/wiki/Heinrich_Hertz
- ThePhysicist 6y agoIf you had asked Erwin Schrödinger how his wave equation could be commercialized he'd probably wouldn't have had any idea either. 100 years later it's the basis for most of our advanced computer technology. I predict the same thing will happen with the standard model, it's just that we can't see the applications yet because we're not technologically advanced enough.
- yyyymmddh 6y agoUpton Sinclair was a socialist. Why should we care what a socialist has to say?
- cmiles74 6y ago
- areoform 6y agoOh wow. Google has become Evil. Isn't this exactly what "Don't Be Evil" was supposed to ward against?
- mrybczyn 6y agoWell, "Don't Be Evil" was a very successful PR slogan. "Framing in the positive" in another context. Once the corporation grows beyond Dunbar's number, it probably begins to look and act just like every other corporation...
- hfkldjsjfkdj 6y agoThis comment is getting old now...
- godelzilla 6y agoAI is marketing, not science. That's called data science.
- zests 6y ago$COMPANY is a private company and they can censor however they see fit. If you do not like it, you can start your own company and do it your own way. If you agree with this statement for one company you should agree with it for another.
- Barrin92 6y agoI don't agree with it for any company. When the private sector crowds out basic research and takes over the function of public institutions, there are three reasonable scenarios. 1. they let the researchers do what they want and science goes on, 2. they're regulated so they can't compromise science, 3. they do what Google does, in which case it's time to take the gloves off and go to 2. because the integrity of scientific research for its own sake is more important than whatever corporate dystopia Google is trying to build.
- zests 6y agoThis take does not make any sense to me. The scientific community is responsible for peer reviews and journals can operate however they’d like. Google is responsible for google and can submit/censor/only pay for research that they want to pay for. Good luck finding any academic that is not beholden to their source of funding. Academia is not the “uncompromised land of science” that it may seem to be. What google is doing is par for the course. I have about 50 other complaints about academia before I get to my complaints about private companies censoring research (which I believe is a very harsh take on the situation at hand.)
- yters 6y agoResponse to the paper arguing many ML models are very underspecified?
- gmkiv 6y agoThe directive to strike a positive tone could be nefarious or benign; we can’t tell from the info in the article. Every research paper has a tone that depends on the personality of the author - I’ve seen it even in my dry field of quantum optics. Every result has many implications and how you discuss them, which implications you emphasize, and the language you use will affect the tone, even if you are not explicitly furthering an agenda. A given article could have a range of legitimate tones, depending on who wrote it. Asking a researcher to be on the positive end of the legitimate spectrum is benign. It can of course go to far.
- WaitWaitWha 6y agoThere are many comments that identify the work by corporate sponsored researchers as fraud and not research. I have the unpopular opinion, and I vehemently disagree. There is no altruistic, sponsored research, no matter of the sponsor, be it corporate, government, non profit, or the Holy See. There is always a specific outcome the sponsor hopes for. The question is, how much pressure the sponsor puts on the researcher not to publish failed or unfavorable outcomes? And, we all know majority of failed research never gets published. Let's presume that we accept sponsorship. What is acceptable level of pressure? I think this is very subjective. To be clear, falsifying by modification, omission or addition of data is fraud. Drawing the wrong conclusion is not. The only time I have seen altruistic, "clean" research was performed by individuals as hobby. Once there are more than one person involved, there will be some preferred outcome. Nature of humanity.
- ska 6y agoI think there is a useful distinction to be made between the impact that financial support has just by existing, and the leverage that can give the sponsor if they choose to use it. If I offer a 1mm grant the use of X drug on Y disease, I am being opinionated about where time should be spent and what diseases are "important". This is not neutral, and can be very political. However, this is in a different universe than if I offer the grant with strings attached like editorial control on the paper(s), or a veto on publishing. As you note, most failed research doesn't get published but vanishingly small amounts of that are due to outside pressure, it's mostly more mundane reasons. Research is sponsored by, and done by people, hence it is flawed. But the best of is places value on the rigor of the research regardless of outcome, and certainly does not attempt to silence results that are not favourable to the sponsors goals. Commercially sponsored work should aspire to this in the same way that public research does - the specific outcome the sponsor hopes for is progress in our understanding. In both cases, sometimes the results are damaging to other projects you have (business goals, policy programs, etc.) which is the real test.
- shadowgovt 6y agoOffering a grant on Y disease at all also biases research in favor of pursuing that over other diseases, possibly diseases of more pressing urgency to cure in the realm of improving human quality of life.
- UncleOxidant 6y agoNo person or company or country sets out to be evil. It happens little by little.
- titzer 6y agoI published several papers while working at Google on Chrome. There is a publication review process that at that time was not onerous at all. In fact, all of my publication submissions were reviewed post facto and basically rubber-stamped. All, with the exception of one, our publication on Spectre, which was carefully vetted to not disclose any sensitive information about further vulnerabilities or violate NDAs with partners. So my experience for publishing at Google, admittedly in a non-controversial area of Programming Languages, was positive. That said, the corporate culture is bending authoritarian and Orwellian. The new changes are just an further evolution in that arc. This is part of the reason I left in 2019.
- googlethrow3134 6y agoWhen have trillion dollar companies not been authoritarian in protecting and furthering their own corporate interests?
- UncleOxidant 6y agoCorporations, companies are some of the last bastions of authoritarianism.
- shadowgovt 6y agoAll too true. The biggest challenge Google is struggling through right now is that it's not just impractical, but dangerous for it to continue to behave like a small and scrappy startup. Behavior patterns that are beneficial for a small company are harmful to a big company. To give a concrete example, "move fast and break things" is fine when you have 100 people relying on you, and should probably be sanctionable (though, in our industry, it is often not outside the context of contracts and pre-arranged agreements) when you have 2 billion people using your system several dozen times a day for critical information, where their lives are materially negatively impacted if that information is false or out-dated.
- rightbyte 6y ago> To give a concrete example, "move fast and break things" is fine when you have 100 people relying on you, If you are OK with things to break you are not relying on it, I would say.
- lacker 6y agoThis whole saga is such a PR “own goal” by Google. The vast majority of corporate research labs just don’t fund research doubting their own ethics in the first place. Company leadership should choose an ethical stance and defend it. It’s an abdication of responsibility to say, well the ethics are an open question, we’re going to fund some research into it. And now Google is being punished for their abdication of responsibility. If they had simply established a corporate policy that said, we believe facial recognition is good when XYZ conditions are met, and defended it, they wouldn’t be in this situation.
- fossuser 6y agoA lot of the stuff really is an open question though. For example if you’re training models on historical data to determine credit worthiness, you can end up with models that penalize people based on race, or zip code (as a proxy for race), or names (same thing again). How to correct for this is not obvious, removing the fields from the data usually just has it appear in some other way that’s correlated to race. My understanding of current approach is to leave the fields in and then check to see how out of whack things are and there are a few approaches to try and correct for this (also an open question). The reason modeling risk on race is unfair is because while the color of your skin may be strongly correlated with being a bad (or good) credit risk due to a whole host of complex historical reasons that lead to bias in the data, it’s not because of someone’s skin color that they should be considered a bad (or good) credit risk. You want the models to be judging based on more causal details. Not data that’s biased due to historical favoritism or exclusion. Doing this poorly can further cement the pattern in the data and make it worse over time in addition to just being less accurate. That said, you want people working on this looking for pragmatic solutions and not just promoting their own agenda while spinning things to seem worse than they are with emotionally loaded language targeting the company. Making demands and then going public on Twitter about your failed ultimatum doesn’t inspire much confidence.
- hfkldjsjfkdj 6y agoThe whole topic around bias in AI is overblown. If anything is to come out to the real world and fails to be inclusive then Google will be swept off their feet by the competition and as long as that is the case then I don't see what I need to be so cautious about. And problems such as political bias run far deeper than some recommendation system. I feel as though these things are just trendy to talk about.
- latenightcoding 6y agoIt is common for university professors to be advisors or C level execs in startups while they still hold academic positions (specially in AI), for me that's many times more concerning.
- bra-ket 6y agoGebru was fired for unprofessional behavior, not because of her research. Jeff Dean’s letter with details: https://docs.google.com/document/u/0/d/1f2kYWDXwhzYnq8ebVtuk9CqQqz7ScqxhSIxeYGrWjK0/mobilebasic https://docs.google.com/document/u/0/d/1f2kYWDXwhzYnq8ebVtuk...
- loopz 6y agoThat can also be true.
- rustytherust 6y agoIt is almost 2021 and the only stuff ai has “solved” is annoying customer support chat bots, endless youtube videos and funny deep fakes. Just like IE it will be 2022 by the time ai will predict covid and 2030 by the time it will find something useful. What happened to them 10 per day ai startup posts in here?