Water required by data centers is increasing considerably with the advent of generative AI. Large language models trained on vast amounts of data need substantial cooling.
While some may argue that water can be reused, it is often pulled from water supplies in times of significant heatwaves, when residents are in need of it as well, and supplies diminsh faster.
In July of 2022, Microsoft pumped in about 11.5 million gallons of water to its cluster of Iowa data centers, according to the West Des Moines Water Works. That’s around 6% of all the water used in the district, which also supplies drinking water to residents.
Meanwhile, the energy required is of course not recycled – and it’s to be expected that energy use is growing as well. The companies behind the tools are just not disclosing how much their AI investments are costing us when it comes to their carbon footprint.
”In its latest environmental report, Microsoft disclosed that its global water consumption spiked 34% from 2021 to 2022 (to nearly 1.7 billion gallons, or more than 2,500 Olympic-sized swimming pools), a sharp increase compared to previous years that outside researchers tie to its AI research.”
https://fortune.com/2023/09/09/ai-chatgpt-usage-fuels-spike-in-microsoft-water-consumption/
Per Axbom
Water required by data centers is increasing considerably with the advent of generative AI. Large language models trained on vast amounts of data need considerable cooling.
While some may argue that water can be reused, it is often pulled from water supplies in times of considerable heatwaves, when residents are in need of it as well, and supplies diminsh faster.
In July of 2022, Microsoft pumped in about 11.5 million gallons of water to its cluster of Iowa data centers, according to the West Des Moines Water Works. That’s around 6% of all the water used in the district, which also supplies drinking water to residents.
Meanwhile, the energy required is of course not recycled – and it’s to be expected that energy use is growing as well. The companies behind the tools are just not disclosing how much their AI investments are costing us when it comes to their carbon footprint.
”In its latest environmental report, Microsoft disclosed that its global water consumption spiked 34% from 2021 to 2022 (to nearly 1.7 billion gallons, or more than 2,500 Olympic-sized swimming pools), a sharp increase compared to previous years that outside researchers tie to its AI research.”
A.I. usage fuels spike in Microsoft’s water consumption | Fortune
Great summary by Dave Karpf of Brian Merchant's "Blood in the Machine".
"The craftsmanship of the book animates his discussion of the modern-day Luddites. Instead of Kirkpatrick Sale’s apocalyptic warnings, Merchant urges us to think seriously about power and inequality today. The modern-day Luddites aren’t rejecting technology and living simply, off-the-grid. They’re people like Chris Smalls and Timnit Gebru, organizing resistance and demanding that the gains from technology ought to benefit workers and communities.
The modern-day Luddites, in other words, are critical technologists. For the most part, they make use of legitimate outlets for resistance (ones that were not available 200+ years ago). They are not opposed to technological progress; they are challenging lazy, self-serving definitions of technological progress."
"You don’t have to reject technology to be a Luddite. That was always a bait-and-switch.
Merchant convincingly demonstrates that a Luddite is someone who thinks about power, and who demands that we build a digital future where we take seriously how the increases in wealth and prosperity will be distributed.
Take it as a call-to-arms: We should all be Luddites now."
https://davekarpf.substack.com/p/we-should-all-be-luddites-now
Why would Nissan want to know about your sexual preferences? Simple: because car brands have found a source of income in selling your data, and your inferred preferences, to third parties.
Mozilla Foundation have recently published their research [1] into how 25 major car brands collect and share deeply personal data, including sexual activity, facial expressions, and genetic and health information.
They do this through car sensors, but also through direct contact with dealerships, their website, social media (yes!), and any of their apps you may download - all together working to build your profile and make assumptions about ("infer"), what your preferences are.
I myself took some time to read Nissan's US privacy notice [2] and it's dismal reading to say the least. Under the heading "Categories of information we collect and disclose", there is this:
"Sensitive personal information, including driver’s license number, national or state identification number, citizenship status, immigration status, race, national origin, religious or philosophical beliefs, sexual orientation, sexual activity, precise geolocation, health diagnosis data, and genetic information."
They can also share and even sell "Inferences drawn from any Personal Data collected to create a profile about a consumer reflecting the consumer’s […] psychological trends, predispositions, behavior, attitudes, intelligence, abilities, and aptitudes" to others for targeted marketing purposes."
This means that one thing they infer from all the data is for example how smart they think you are.
Of course, if you are a passenger in a Nissan you will also be subjected to monitoring by sensors, and Nissan say it's up to the owner to tell passengers this surveillance is happening.
Oh, you think this is only about Nissan? Think again.
I haven't had a chance to see to what extent these practices are happening in the EU (where many of them of course are illegal), but the fact remains that there is a ton of surveillance going on in a modern car and you may want to start thinking about where that information ends up, how it is stored, who has access to it and how it is used.
[1] ‘Privacy Nightmare on Wheels’, research summary
[2] Nissan USA privacy notice
https://www.nissanusa.com/privacy.html
[3] What Data Does My Car Collect About Me and Where Does It Go
#DigitalEthics
First season done and I have to admit I absolutely love One Piece on Netflix. In my case I didn’t know what to expect but the characters and the live action are right up my alley.
To be clear, I was not familiar with the anime so there are no adaptation quirks for me to get hung up on.
#OnePiece #Netflix
”The way of life can be free and beautiful, but we have lost the way. Greed has poisoned men’s souls, has barricaded the world with hate, has goose-steped us into misery and bloodshed.
We have developed speed, but we have shut ourselves in. Machinery that gives abundance has left us in want.
Our knowledge has made us cynical. Our cleverness, hard and unkind. We think too much and feel too little. More than machinery we need humanity. More than cleverness we need kindness and gentleness.
Without these qualities, life will be violent and all will be lost.”
After two years of filming the script he put together in 1938, The Great Dictator premiered in October of 1940. It was British comedian Charlie Chaplin’s first speaking role. At the time of its release, the United States – where Chaplin lived – was still formally at peace with Nazi Germany and neutral during what were the early days of World War II. Chaplin's film helped build a strong condemnation of the German and Italian dictators Adolf Hitler and Benito Mussolini, as well as fascism, antisemitism, and the Nazis.
In his 1964 autobiography, Chaplin stated that he could not have made the film if he had known about the true extent of the horrors of the Nazi concentration camps at that time.
https://www.youtube.com/watch?v=J7GY1Xg6X20&list=PLyU6zePFLJ-DJ62sKt89bh0l9zOx0btI-&index=1
@Midgard Men herrejävlar vad illa!!! Och ljuger dessutom om att ha fått uppgifter av andra...
@chromatic This is extra funny because I had this exact conversation with a friend I shared the story with.
“According to Mozilla research, popular global brands — including BMW, Ford, Toyota, Tesla, Kia, and Subaru — can collect deeply personal data such as sexual activity, immigration status, race, facial expressions, weight, health and genetic information, and where you drive.
Researchers found data is being gathered by sensors, microphones, cameras, and the phones and devices drivers connect to their cars, as well as by car apps, company websites, dealerships, and vehicle telematics.
Brands can then share or sell this data to third parties. Car brands can also take much of this data and use it to develop inferences about a driver’s intelligence, abilities, characteristics, preferences, and more.”
https://foundation.mozilla.org/en/blog/privacy-nightmare-on-wheels-every-car-brand-reviewed-by-mozilla-including-ford-volkswagen-and-toyota-flunks-privacy-test/
I did some further reading after sharing this, and it turns out that ”sexual activity” is actually mentioned verbatim in the Nissan USA privacy notice.
Fromt Nissan’s very real Privacy notice, under ”Categories of information we collect and disclose”:
”Sensitive personal information, including driver’s license number, national or state identification number, citizenship status, immigration status, race, national origin, religious or philosophical beliefs, sexual orientation, sexual activity, precise geolocation, health diagnosis data, and genetic information.”
Not only that… as Mozilla Foundation writes:
”They also say they can share and even sell "Inferences drawn from any Personal Data collected to create a profile about a consumer reflecting the consumer’s preferences, characteristics, psychological trends, predispositions, behavior, attitudes, intelligence, abilities, and aptitudes" to others for targeted marketing purposes. Yes, Nissan says they can infer things like how smart you are, if you have a predisposition to drink, if you are acting depressed, and if you are any good at chess (we're guessing that's what they can infer..it could be even worse than that), and then they say they can make as much money off that very personal information as they can. Nissan, you suck.”
https://foundation.mozilla.org/en/privacynotincluded/nissan/
Some of you are wondering how the cars collect and share this if they have no internet connection. A lot of this data is actually collected through other means, and when you are in touch with a dealership. So it’s direct contact but also info they proactively collect through social media (not kidding! I just read Nissan’s privacy notice again) and credit reporting entities. If you have downloaded one of their apps, the internet connection is right there.
Of course, a lot more cars than you imagine have internet connections, and cars have had some sort of onboard computer since the 1970s. A lot of data is stored until it can be accessed or uploaded. And you often don’t even have to press buttons for something to be logged. Sensors are always on, marketed as making you safer, but also saving data to be sold to third parties.
So car companies may also combine information collected about you from your car with personal information they get from third parties. Then they can share (or even sell) that information, and any “inferences” they made based on it, to all kinds of businesses
And here’s another kicker… just by sitting in a vehicle that uses NissanConnect services, you agree to have your data collected by Nissan. So if you hitch a ride with a friend’s Nissan, you are on Nissan’s radar. The privacy policy makes it the responsibility of the owner to disclose this to anyone travelling in their car.
https://foundation.mozilla.org/en/privacynotincluded/articles/what-data-does-my-car-collect-about-me-and-where-does-it-go/
Hur försiktig är du när du ska skanna en QR-kod? I mitt senaste inlägg berättar jag om hur ett hotell jag besökte i somras lät lappar med QR-koder spridas som konfetti på borden i lobbyn. Hotellet som avsändare känns tryggt, men någon av alla människor som passerar borden kan förstås byta ut en eller flera av dessa lappar mot sin egen… och jag leker med vad det scenariot kan innebära.
Vad QR-koden var till för? Att boka tid för frukost. Det låter jag förstås inte heller gå okommenterat.
Med denna text börjar jag hitta tillbaka till en ton i mitt språk som jag njuter helhjärtat av att skriva.
https://axbom.se/hotell-qr-kod/
"More than once, she challenged Bill Gates face-to-face to fix problems with Internet Explorer."
An absolute legend. This is such sad news.
#MollyHolzschlag #Accessibility #WebStandards #a11y
Hur försiktig är du när du ska skanna en QR-kod? I mitt senaste inlägg berättar jag om hur ett hotell jag besökte i somras lät lappar med QR-koder spridas som konfetti på borden i lobbyn. Hotellet som avsändare känns tryggt, men någon av alla människor som passerar borden kan förstås byta ut en eller flera av dessa lappar mot sin egen… och jag leker med vad det scenariot kan innebära.
Vad QR-koden var till för? Att boka tid för frukost. Det låter jag förstås inte heller gå okommenterat.
@fbaum Yeah, I presume you're thinking of a situation where people don't introduce themselves and there is no way to ascertain that a bot is participating?
@GuyDudeman
Google Duet is the most talked about. But we also started getting pitches for CEOs of companies that make these to be on our podcast…
https://www.businessinsider.com/google-duet-ai-attend-meetings-workplace-productivity-2023-8?op=1&r=US&IR=T
@thomkennon Of course. I'll likely write a longer-form post on my blog about it if you want to wait on that. But. go ahead if you want :)
Heads up, because this will likely start happening more regularly in the next six months:
If a person sends a bot (will be marketed as an ”AI assistant”) to participate in their place at an online meeting you are hosting, your safest option is to kick it out. That bot will be recording everything and using it for things that meeting participants have not consented to.
Updates:
Google Duet is the most talked about at the moment, but we also started getting pitches for CEOs of companies that make these to be on our podcast…
https://www.businessinsider.com/google-duet-ai-attend-meetings-workplace-productivity-2023-8?op=1&r=US&IR=T
Yes, I know recording meetings can aldready be a problem but this will certainly amplify, and perhaps normalize, the practice - I also wrote about issues with asking consent too late in meetings, here: https://axbom.com/consent-for-recording-meetings/
So even if the bot asks consent (by voice?) as soon as it joins a meeting, this won't be enough in many cases. It would also have to disclose subprocessors and data management practices. Imagine starting a meeting with this and each participant having to respond 😬
Anyway, this was just my observation and take on it. From a data protection perspective you have responsibilities as a meeting host if you know it's being recorded, hence why I wrote "you are hosting".
And yes, I am in the EU.
@sam_c Perfectly put 🙌
I love writing and reflecting.
Writing makes me happy.
Reflecting helps me wrap my head around complexity. And write better.
Using general-purpose language models to write or reason makes me feel queasy and sad.
For me it's as clear-cut as that.
I'd rather feel happy.
I'd rather reflect on terms I can control and where the source of information is transparent.
And honestly, I'm okay with others feeling differently about their own experience.
As long as everyone gets to choose for themselves.
As long as other people aren't being hurt in the process.
@stephenpa I do like playful names, and I’m intrigued but also a tad skeptical… would that be contrasted with prototypes that are designed to *not* provoke a reaction/conversation? And would the reaction be the purpose rather than learning something about product fit and understanding?
So many questions 😅
@EFreethought
Not a bad one. I'd add that to possible and plausible. But not preferable 😉
@failedLyndonLaRouchite