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 perspectives on their employer's products."
Like... these AI researchers are delusional if they think their work is truly unbiased and that they can work at a company and get paid by that company to produce unbiased research. If I took that job at Juul even though it was ridiculous $$ I would have known what I was getting into and who was buttering my bread.
It's either sell your soul and move to industry and enjoy the industry $$ or stay in academia and enjoy your moral purity and being able to look yourself in the mirror at the end of the day but not being able to buy a house in the Bay Area.
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.
I have a very hard time believing you can retain much moral purity after all the publish, publish, publish, grants and will they or won’t they tenure drama. At least the motives of the company hiring you to do research for them are clear.
It's just a bitch that for-profit company lobby for no regulations, and fire people who speak up to power.
However, we all know company culture do not tolerate speaking up to power. So these people also went about it the wrong way, and going public too soon over something that was just theoretical.
What you need to do is gather evidence, and escalate as security violations internally. Use the proper channels. This is harder work than crying wolf, obviously.
This is incredibly spot-on way to describe navigating through delicate office channels to create change and also not ruffle too many feathers. Well said! Can you sit on my shoulder please and advise me real-time throughout my career? Comments like this are why hacker news is so much better than reddit!
It seems that the issue here is that Google didn't hire dispassionate ethical researchers able to deal with a wide range of ethical issues; instead it hired activists with a precise, narrow agenda. People who think that the world itself is biased (including the company they work for and their own management) and that Google should change its products to enact a change upon it. Once these activists have found themselves cuddled the most powerful company in the world and showered with praise and money, it seems their egos just started inflating until they burst.
My wish is for companies to stop outputting "research" and researchers to work for the correctly aligned institutions...and accept the pay cut that comes with it.
AI has at least as much potential to be dangerous, ethical constraints on it are important when you consider the failure of common sense and community good will for SN and BD. However, every berk going latched onto the idea that they could write some sort of rubbish about "applying type 1 and type 2 thinking in AI design" or "why robots can't really be people" and have founded community dedicated to promoting themselves and getting money at the expense of insight and real work and real constraints developed by people who know what they are on about and can really help.
You are right about the humanities types - the otherside of the coin is that there is a definite class of insight that should come out of social sciences and history that is relevant here. Unfortunately those fields are dominated by bullshit artists which means that the signal to noise ratio is very very low.
But i agree that many of the people who have such job titles, and make it into the media, seem to be wingnuts
You can be as cynical as you want but the fact Google fired them with little rhyme or reason and destroyed their team is BAD thing, we should be upset at Google for it.
I hace no expertise in the field of AI ethics, and I'm not sure what to think about the Timnit drama in general, but I don't think what you're saying here is really accurate. There can be a lot of value in pointing out problems even if you're not the one who can provide a solution. If no-one else has spotted the problem then pointing it out can draw the attention of those who might know how to fix it before things get worse.
If you can say "hey, there's a flaw in the design of this rocket that means it'll explode if we launch it, wasting millions of dollars" then that's valuable to your company, even if you don't know how to design a rocket that works.