Guest essay by Eric Worrall
When the woke outwoke the woke. Back in December, Google fired AI Ethics Unit co-leader Timnit Gebru, in relation to her paper “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?”. Google have now just fired their other ethics head, Margaret Mitchell, apparently for trying to gather evidence while investigating the ousting of Timnit.
I’m fired: Google AI in meltdown as ethics unit co-lead forced out just weeks after coworker ousted
Plus: IBM reportedly trying to sell Watson AI Health, and more
Katyanna Quach Mon 22 Feb 2021 // 12:21 UTC
Google has finished its probe into the controversial ousting of Timnit Gebru, co-leader of its Ethical AI unit. The ad giant promised to implement new procedures around “potentially sensitive employee exits,” though it did not make its findings public.
Gebru said she was fired for warning coworkers in an internal memo that, due to management apathy, it was a waste of energy trying to foster diversity, equality, and inclusion within the Silicon Valley goliath. Google claimed she effectively resigned.
…
Meanwhile, Margaret Mitchell, who also co-led the Ethical AI unit alongside Gebru, said on Friday she has been fired. Mitchell had been locked out of her corporate account for weeks.
…
Read more: https://www.theregister.com/2021/02/22/in_brief_ai/
From the paper “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?”;
…
3 ENVIRONMENTAL AND FINANCIAL COST
Strubell et al. recently benchmarked model training and develop- ment costs in terms of dollars and estimated 𝐶𝑂2 emissions [129]. While the average human is responsible for an estimated 5t 𝐶𝑂2𝑒 per year,2 the authors trained a Transformer (big) model [136] with neural architecture search and estimated that the training procedure emitted 284t of 𝐶𝑂2. Training a single BERT base model (without hyperparameter tuning) on GPUs was estimated to require as much energy as a trans-American flight.
While some of this energy comes from renewable sources, or cloud compute companies’ use of carbon credit-offset sources, the authors note that the majority of cloud compute providers’ energy is not sourced from renewable sources and many energy sources in the world are not carbon neutral. In addition, renewable energy sources are still costly to the environment,3 and data centers with increasing computation requirements take away from other potential uses of green energy,4 underscoring the need for energy efficient model architectures and training paradigms.
…
Read more: http://faculty.washington.edu/ebender/papers/Stochastic_Parrots.pdf
Timnit has also criticised other issues with AIs, for example in 2018 she helped stop the rollout of an Amazon facial recognition system being used by police agencies, by demonstrating the flawed Amazon system was 34% less capable of correctly identifying black women, compared to its ability to correctly identify white men. The problem – the dataset used to train the AI mostly contained white faces.
What can I say – losing one ethics head could be an accident. Losing two in quick succession starts to look like carelessness, perhaps even raises suspicions that what Google really wants is a compliant ethics team which does whatever top management tells them to do.
A heretic of an ostensibly “secular” quasi-religion.
What else to expect from an organisation that supports global dictatorship based on no tangible CO2 effect evidence.
They supported a global dictatorship long before anyone thought of blaming CO2 for everything.
Well, if she did not learn her lesson – next stop is total cancellation – or bye bye…….also known as Google Gag.
Sent to the Guglag?
Perhaps the Googlag?
Will no one rid me of this bothersome priestess… ethicist.
LOVE IT…..LONG LIVE THE KING!!!!!
… but Henry II died of a bleeding ulcer
Yeah… s-u-u-u-r-e… just keep believing that.
😜
According to Sir David Attenborough, the bleeding ulcer was caused by Henry’s worry about climate change.
Compliance would be the word of the day and what they are after.
“I’m sorry, your Gargle search could not be completed because it is not sunny enough at the server farm.”
You’re social credits have been deemed to be no longer viable. Throw another baby on the barbie, cannibalize her profitable parts, and sequester her carbon pollutants, NOW.
all I have to say is LMAO
Useful idiots never do figure out what will happen to them once they are no longer useful.
Which is why they are useful in the first place.
and why they remain idiots
“Google claimed she effectively resigned.”….So she was asked to consider the consequences of her NOT receiving any more paychecks after Friday……
Why not? All the low hanging realist people have been cancelled so what else is a hateful organization to do but eat their own? We are at the beginning of Progressive self immolation and it will only get better from here. No one, no one, will be aware enough to pass the woke litmus test and they’ll all be looking over their shoulders for people trying to take their place. The circular shooting gallery is gaining steam. AGW is poised to be outed by its’ progenitors as they realize they’re next on the hit list.
“eat their own”… never changes, all disposable, all Gulagable.
These people need to read Solzhenitsyn, they might begin to understand their situation a bit clearer.
Great author.
I laugh so hard when this happens as once upon a time these were the same ones who opposed big business/government and book burning and censorship and now that their party is in power…Surprise surprise. Too bleeping funny!!!!
The further we twist & turn to ensure every aspect of human life conforms with “low carbon output”, the closer we are to disappearing up our own fundaments.
Boo frickin’ hoo.
It’s a matter of principle… or, alternatively, em-pathetic appeal.
If I ran Google, I’d have fired her too. She’s a fool to have believed that her job had any meaning.
Of course they do. Google is no different in that than any other large corporation. The “ethics team” is nothing but politic virtue signalling for the brain dead media. Corporations are only interested in their own best interests. There is no room for altruism in the board room.
Google leads the way with ‘cancel culture’ for a reason. Why would anyone think they weren’t prepared to do it internally?
Can’t go wrong paraphrasing Lady Bracknell, especially when it works so well.
“A Hand bag?”
They are progressives after all, and that is how progressives operate.
I love it when the Left eat their own. It makes my day.
It’s pretty bad when Google is banned in China because the CCP hates competition.
Pass the hot sauce. Reminds me of an old joke, the punchline to which is, “Does this taste funny to you?”
Guess the lesson is, Don’t Question the Kool Aide.
Maggie should have followed in the footsteps of her namesake, and taken up writing a contemporary version of “Gone With The Wind”
Maybe it could be called “Gone With The Sinned”?
But frankly Mr., Google doesn’t give damn.
I’ll think about that tomorrow.
Perhaps just “Gone”.
Google is an IT company and everybody who works there understands that running computers as in performing computations takes energy.
With regard to this AI effort, once the Board of Directors, through whatever process and proxies are involved, approves the project then the consumption of resources (human, financial and computing) is approved as well.
It would be an “ethics” issue if the project leaders misrepresented the costs, benefits, or any other parameters that were used to determine whether, or not, the project should proceed.
It would be an “ethics” issue if the intended use of this AI was going to be used for immoral and/or illegal purposes. You get the idea . . . .
Thought experiment . . . . If the cost in computing resources were, say, 1% of what the ethics group observed would they object?
If the answer is “no” then IMO they’ve definitely not raised an “ethics” issue. As I read the piece, they’re not objecting to the effort or the way it is being carried out. Rather, they’ve determined that it costs too much, in terms of CO2 even though management already approved the work.
In other words, looks to me that they’ve strayed outside of their lane. I can’t think of any company that would build and maintain an ethics oversight organization to question the very basis of their existence.
Fundamental rule in business . . . stay in your “lane” or you might get fired.
This is quite old in Internet timeframes, they just fired another one
https://finance.yahoo.com/news/google-fires-second-ai-ethics-022428541.html
You shouldn’t be too happy about the left eating themselves. These are the useful idiots. The machine that creates them chugs along because power is in the hands of the media and teachers, and the only people with power to warn the people are the media and teachers.
Most Americans wouldn’t be aware that our PM in Australia is getting extremely bad press for an alleged rape of a staffer – by another staffer in the office of another minister that he never heard about because she chose to not to go the police. Our PM was attacked as if he was the accused for almost a week before she went to the police. She claimed that she didn’t go to the police because she thought it would affect her career. Strangely, the media accepted this even though she went to the press and held off going to the police for nearly a week afterwards.
Our PM is still copping it even from supposedly rightwing press (ie press that don’t censor everything from a conservative) because an election is due soon.
During the last election, the media went mute when allegations of rape were made against the opposition party leader. Justice was allowed to take it’s course but the police decided that there was no corroborating evidence. Soon after (There was a law change) charges were made against priests, including Cardinal Pell, with only a single witness, the accuser and alleged victim, as evidence. Both were convicted but these were overturned because evidence that raised significant doubt about how credible the testimony was, was ignored. Most of the media described it as a technicality.
We’ve passed a tipping point that has nothing to do with climate.
Yes and the way the Metoo crowd piled on a female Minister who essentially handled the victim’s claim according to their wishes beggared belief. Apparently the Minister should have frogmarched her down to the police station whilst notifying the PM and leaders of all other parties that national affairs had to stop forthwith while they all came together for a perfect solution to see such an incident never happened again to anyone remotely connected to parliament. With appropriate media release of course. Otherwise this was a cover up.
The Minister has to resign immediately for not seeing all that clearly but apparently we need more women in politics.
Yep. They ignore that the minister can’t go around accusing someone of rape if the alleged victim does not want the police to investigate. Somehow, common sense becomes a cover up.
Wasnt it the other way around… she was told that a police investigation near to an election could damage your career…but its your choice, promotion and payrise or the victim scrapheap
No such thing but the common mob rule conclusion without any evidence is that’s the way Ms Higgins would/should feel at the time and why she didn’t want a police prosecution. In fact the meeting with the Minister the following week after the Friday night episode was all about the breach of security at which she was told the male staffer had been fired for it. Obviously she was in trouble too. Only then did she make the allegation of rape to the Minister and we can all speculate as to why she hadn’t prior to that meeting.ie first thing Monday morning.
yup then throw in a near schoolie? who claims rape by the same bloke
so why didnt she OR her parents press charges either
its very smelly all round
somethings VERY OFF!
Simple answer to that
“The board, [at AFP]chaired by deputy commissioner Ian McCartney, considered the case five days after Ms Higgins contacted the police about a complaint she had first lodged in April 2019.”
First lodges in April 2019
Here in the US, when a number of women accused Biden of improper behavior, the media covered it up and attacked the women instead.
The paper is one big whine. For example, 64% of users on Reddit are men, therefore training AI models with data crawled from Reddit is sexist. I don’t blame Google for firing her. I wonder what her salary and benefits were. And did anyone even notice when she was gone?
I challenge your logic. If a significant minority of the population get minority representation in a statistical study then there is a danger that the study’s results will be skewed against them. Timnit Gebru is claiming that in this particular case, it has. Her claim should be treated on its merits, and it certainly does seem to be the case that facial recognition is not as good for black women as it is for white men. Personally, I would prefer not to be recognised by a camera in the hands of people that I do not know I can trust, but those cameras can also be good at disproving false accusations. If we have ti have them, then let’s ensure they work properly.
Right, but it sounds like she was fired because of that internal memo accusing Google of being apathetic towards diversity.
So. With these two firings the Gargle admits it no longer has ethics? Not that it ever did. I mean, if I were bent on all encompassing evil I too would claim as my motto; “Don’t be evil“.
Not anymore.
Alphabet dropped it in 2015 https://time.com/4060575/alphabet-google-dont-be-evil/
Apparently Google dropped it in 2018 https://gizmodo.com/google-removes-nearly-all-mentions-of-dont-be-evil-from-1826153393
Google is a bloody evil empire and I have some experience of that.
It is time that silicon valley built churches and statues to Stalin and Mao.
Coming from the company that removed “don’t be evil” as a guiding statement…
Verging OT but interesting..
It does mention Google and AI an awful lot.
and Tesla. and Apple.
Would it be some part of why Google AI is getting ‘a bit fractious’
Tesla’s data advantage. Can Apple, or others, keep up?
Keep in in mind that AI is like a human baby. AI is coded to learn and must be “trained” as this article mentions. That means the AI ultimately reflects the methods and ideas of its trainers. If you have bunch of leftist outright Marxists doing the training then you are going to get “algorithms” that judge in favor of Marxist thought and against anything else. The results are, of necessity, “fractious” by more than just “a bit”.
I think this shows that internally Google management understands “Go woke, go broke,” as the reason for the ‘apathy’ in mid and upper management towards not forcing a diversity outcome as those fired, “woke” idiots wanted.
In any competitive private business, a model driven by profit motive and share price, for it to succeed, it must be a meritocracy on promotions. You would have to be a complete moron manager to promote internally on diversity (skin color, LGBTQ status, gender declarations, etc) rather than the merit of hard work.
A semi-woke HR can help make sure diversity in new-hires are available during recruiting processes on campuses. But once those people are on board, for the company to succeed it must be on merit. If it is not on merit, and promotions decided on skin color or LGBTQ declarations, then the hard workers who get passed over for promotions and bonuses that comes with them,will leave or they will stop working hard. Everything will atrify under a “woke” system in the private sector.
The public sector though is another matter…. There is no competition there, so it takes only 1 shovel and 5 guys to dig a ditch if a municipal crew is put on a small ditch job.
Are Google making way for a position for Peter Gleike ?
OK then.
Calculating the energy it takes to train an AI, based on today’s technology is the wrong thing to do.
Calculating the energy it would take to run a smart phone based on 1950s technology is wrong. As far as I can tell, no 1950s computer would have been capable of running a smart phone and, if they were, they would have taken the energy of a decent sized power plant.
That person may have deserved to be fired for shortsightedness.
The most efficient computer I’m aware of is the human brain. How much energy does it take to train a human to achieve expert performance? Suppose it takes ten years.
Given an input of 1000 food calories per day for ten years, I calculated the energy it takes to train an expert as about 4300 kwh. At a around 20 cents per kwh it takes about $1000 worth of energy to train an expert.
Suppose that a human expert can earn your company $100,000 per year in profit. If you could train AI experts for $1000 to do the same thing, the payback would be less than a day.
Technology will improve. AI will become more efficient. AI energy costs will eventually become irrelevant compared with all the other ethical issues that AI will raise.
I hope that person was fired for stupidity.
“Given an input of 1000 food calories per day for ten years,”
That will be one skinny dude.
How about a petite dudette? 🙂
I thought he was talking about the calories needed by the brain.
Don’t encourage them – there is already a push to put living brain tissue in bottles, and call it an “AI”.
https://www.newscientist.com/article/dn6573-brain-cells-in-a-dish-fly-fighter-plane/
the ship who sang book on exactly that anne mc caffrey ages ago now
Reading this sounds like this human was fire more for questioning the management’s commitment to diversity and equality and publicly commenting than this human’s paper on tonnage of CO2 emitted.
Now this human’s paper was dumb, stupid and unnecessary for a modeling project. There’s a piece of equipment attached to every building that is connected to an electrical grid and has electricity delivered to them from the grid. Not to mention all power monitoring devices in servers frams. Take a reading before you start your model processing; take a reading after your model completes it processing phase. This equals power consumed.
Calling Peter Gleick. Urgent message for Peter Gleick!
An alternative hypothesis is that gender and ethnicity were given more weight in the hiring decision than competence. Perhaps the AI Ethics Unit just realized that.
This reminds me of the Canadian Content rules for Canadian broadcasters.
Darn.
This reminds me of the Canadian Content rules for Canadian broadcasters.
There was a time when most of the music played on Canadian radio stations was pop music from the ‘States. Canadian artists, for whatever reason, didn’t stand a chance of getting on the air.
Was it because Ameryican music was superior? Some of it clearly was. On the other hand, most of the music on the radio was crappy.
Someone in the Canadian government observed that our crappy music was every bit as good as crappy American music.
I will riff on that and say that the crappy third rate work of anyone of any sex and gender and ethnicity is every bit as good as the third rate crappy work turned out by most male Caucasian scholars.
To mock Canadian Content rules are why Dave Thomas and Rick Moranis created the MacKenzie brothers and their “Great White North” show.
Cooorooka coocacoo
Cooorooka coocacoo
eh?
Loved that skit on Second City! And Boys’ In The Hall! And all the great Canadian bands (The Guess Who, BTO, Tragically Hip, Bryan Adams, Blue Rodeo, etc). Judging from that success and lack of suckiness, the Canadian Content Rules tried to fight the herd mentality in pop culture – start with a band that finds some success and pretty soon it goes viral and it is saturation bombed everywhere over and over by radio stations. That’s bad enough inside the States itself, where other good bands might as well take a break from touring while the new fad hogs the limelight. But it’s even worse for those outside the States.
If you assume that good musicians are randomly distributed throughout the population, then the odds are that a population that is 10 times larger will have 10 times as many good musicians.
The market for music, and many other things, is described by a Pareto distribution.
The only irony of note is combining Google and ethics in a single sentence.
The horror…the horror…
Nothing to see here folks move along-
Google Engineers Explain Why They Stopped R&D in Renewable Energy | Greentech Media
Google don’t need know steenking facts as there’s a new administration in town.
There are no good guys here. All of them should lose.
She could try doing some honest work now
But beyond that, we call on the field to recognize that appli- cations that aim to believably mimic humans bring risk of extreme harms. Work on synthetic human behavior is a bright line in ethical AI development, where downstream effects need to be understood and modeled in order to block foreseeable harm to society and different social groups. Thus what is also needed is scholarship on the benefits, harms, and risks of mimicking humans and thoughtful design of target tasks grounded in use cases sufficiently concrete to allow collaborative design with affected communities.
Now turn the job over fully to the computers to automate the process of splintering society into Planck lengths.
First google has only suppression of truth an dissent as their cor ethics.
Second this isn’t AI. Its a kludge of if then what if with only the programmers intelligence.
Artificial…
but NOT intelligent !
Google is evil, period, end of story.
Eventually they WILL COME FOR YOU TOO
yeah cloud n other servers energy use is massive
as is the water use for cooling
funny how sensitive they get when thats raised isnt it?
It requires a total or near total ban. That is the ethical solution.
Ethics committees don’t exist to promote ethical behavior, they exist to justify the decisions already made elsewhere. Medical ethics boards are prime examples. “It’s ethical to remove food and water from this person because their quality of life is so poor.” “It’s ethical to experiment using human embryos because of the life-saving potential.”
Google expects the same thing from their ethics team. “CO2 is bad, but our increased emissions is ethical because we’re using it to improve society. Everyone else, though, should emit less, those uncaring bastards.”
They want a rubber-stamp of “this is ethical”, not an actual look at the ethics of their practices.
Losing two in quick succession
======
Google fails to recognize the Importance of Being Earnest.
“Timnit has also criticised other issues with AIs, for example in 2018 she helped stop the rollout of an Amazon facial recognition system being used by police agencies, by demonstrating the flawed Amazon system was 34% less capable of correctly identifying black women, compared to its ability to correctly identify white men. The problem – the dataset used to train the AI mostly contained white faces.” What a racist organisation! even in the artificial “intelligence” sphere.
Web site like this make mokery of oppressed peeples. me native tong is Tlingit and haff time kant read. because need trnslation sevices. Tlingit peeples need to rize up for they rites.
How you like need to go in Tlingit-speaking countries, not no where the man’s room is? Then show is on other feet!
Tlingit need equal rites.
White man come, kill our women, rake our buffalo, now this… warming take our icees. where the great narwhal hunts of yesteryear? Gone!
Having an ethics leader is just virtue signaling. Fine so long as they dont mess with the $$$ or question upper management
AI is already voting for you
The proles are restless-
Exclusive: Google pledges changes to research oversight after internal revolt (msn.com)