Essay by Eric Worrall
According to advocates, AI can generate better flood predictions than physics and geography. But is this just more magic box thinking?
Should AI’s Role To Cut Greenhouse Gas Emissions Be Greater?
22 hours ago Carolyn Fortuna
Scientists warn that heat waves, floods, droughts, and severe storms will get far worse in the decades ahead unless we change course. Looking ahead, could AI’s role in developing new climate models save us many gigatons of carbon emissions?
…
AI’s role in the struggle against climate change is already prominent and is also controversial. While it seems evident that AI can serve in the pursuit of a greener future, checks and balances that ensure fairness and equity must be implemented.
…
For decades, scientists looked at climate prediction models based largely on the rules of physics and chemistry to forecast weather patterns. Now hybrid-based models consider machine learning and other generative AI tools which help climate scientists create even more accurate and precise systems. For example, doctoral students who are working with officials from the Tennessee Valley Authority to provide a more accurate hybrid-based flood prediction system than the one they are using that is based solely on physics.
…
Read more: https://cleantechnica.com/2023/12/31/should-ais-role-to-cut-greenhouse-gas-emissions-be-greater/
The article also provides a link to their ICEF submission;
ARTIFICIAL INTELLIGENCE FOR CLIMATE CHANGE MITIGATION ROADMAP
Innovation for Cool Earth Forum
DRAFT FOR COMMENT
October 2023…
PREFACE
Artificial intelligence (AI) is a hot topic. One business leader recently called it “the defining
technology of our time.” Another said “It is difficult to think of a major industry that AI will not
transform.”Meanwhile countries around the world are struggling to respond to the challenge of climate change. Despite encouraging developments, including steep declines in the price of renewable power, global emissions of greenhouse gases keep rising. Scientists warn that heat waves, floods, droughts and severe storms will get far worse in the decades ahead unless we change course.
Can AI help cut emissions of greenhouse gases? This roadmap explores that question. Our goal is to provide a useful resource for experts and non-experts alike. In Part I of the roadmap, we provide
brief introductions to both AI and climate change. In Part II, we explore six areas in which AI is
helping respond to climate change and could do much more. (These are greenhouse gas emissions monitoring, the power grid, manufacturing, materials innovation, the food system and road transport.) In Part III, we explore cross-cutting barriers, risks and policies. We finish with findings and recommendations.The relationship between AI and climate change is a big topic. Among the questions we do not
explore in this roadmap are (1) how AI could contribute to climate change adaptation (an important
area for work and study) and (2) whether the broad societal forces that AI may unleash are more
likely to help or hinder the response to climate change (a difficult question in light of the many
uncertainties with respect to AI’s impacts in the years ahead). Instead, we aim to provide a resource that will make favorable outcomes more likely, pointing toward ways in which AI can contribute to climate solutions.This roadmap builds on the body of literature produced annually in connection with the ICEF
conference. Previous roadmaps have addressed the following topics:Read more: ICF AI Roadmap
- Low-Carbon Ammonia (2022)
- Blue Carbon (2022)
- Carbon Mineralization (2021)
- Biomass Carbon Removal and Storage (BiCRS) (2020)
- Industrial Heat Decarbonization (2019)
- Direct Air Capture (2018)
- Carbon Dioxide Utilization (2017 and 2016)
- Energy Storage (2017)
- Zero Energy Buildings (2016)
- Solar and Storage (2015)
The reference to checks and balances is entertaining. From the full report:
… Bias-related risks when using AI for climate mitigation include using AI models that prioritize certain groups due to historic data availability. For example, data for wealthier nations and neighborhoods are often better than data for poorer ones. Privacy-related risks include unauthorized data leaks to third parties, personal identification and even surveillance. Security-related risks are especially acute if AI is used for real-time decision-making (for example in operating factories or the electric grid). …
Read more: ICF AI Roadmap
Chapter 10 Risks explains these bias-related risks in more detail. The risk includes cultural biases, programmer biases and data biases (e.g. putting more solar panels into an area already rich with solar panels, because it is obviously a good place for solar panels – while ignoring other potentially useful locations).
I suspect this point about checks and balances has been added because some of their preliminary AI model runs produced some embarrassing recommendations.
For example if you were to feed an artificial intelligence RCP8.5 climate scenario assumptions, then ask the AI to maximise economic production in a global hothouse scenario, the AI might recommend ignoring renewables and maximising fossil fuel energy production.
If you then program the AI to give more priority to the alleged climate harms to nations like Bangladesh and Arabia, the AI might recommend ignoring renewables, but subsidising air conditioners and building sea walls and flood levees, rather than the politically acceptable recommendation of more climate action.
It would obviously be unacceptable for the AI to produce a product which embarrasses its political backers, so there would likely be a strong temptation for AI researchers to add “checks and balances” to their system until it produces the right answer. A lot like the policymaker review of IPCC reports, or allegedly dubious adjustments of temperature records, except AI scientists who yield to this temptation to prioritise political correctness would be more likely to try to restrain their systems by tweaking the software rather than editing the final product.
Whether this “checks and balances” constrained AI product has any practical value is a different question.
I think any AI would reflect the training documents. GIGO.
Well said, Tom. All AI can do is synthesise from a hodge-podge of various peoples thoughts as found on the internet. It can’t think for itself, so don’t expect any original thinking.
I don’t know. My own experience with “AI” showed that they will just make stuff up, and not show you any references at all to backup their claims. None. In other words, no better than any alarmist.
AI will probably become a thing in “climate science”. Since most people seem to be aware that computer models are largely BS, they’ll try AI and hoodwink a few more credulous fools.
“credulous fools” – Is there any other kind?
Yes, incredulous fools who will not change their position despite persuasive evidence.
So instead of torturing the data to deliver the “correct” answers, climate “scientists” are now going to torture the AI algorithms?
“AI” looks like “Al Gore” in sans serif type.
Explains his boring, wooden almost mechanical way of talking….
Believe it or not, ChatGPT played a major role in enhancing my understanding of measurement uncertainty or better: the ‘Great Unknown’ issue. It’s very useful; you just need to be cautious about discussing uncertainty in specific contexts. The algorithm is deigned to fall back on IPCC talking points when sensitive terms like ‘global warming’ or ‘climate’ are mentioned. Assuming AI intelligence reflects mainstream climate science and the IPCC, this feature greatly assists comprehending the issue and make for much more straightforward debates with climate alarmists.
I had no luck at all with ChatGPT. It basically became an echo chamber back to me as I told to check different sources. It started saying something like “I’m sorry, you are correct that your reference indicates uncertainty should be calculated as you said.”
My conclusion is that it has no intuition and can not process mathematic processes. It can only process the equations you give it but can not understand context or what is to be accomplished.
It “talks” well, but it is no teacher. It speaks more to the nations education system and the inability to write coherently. It will never have a “lightbulb” moment where something NEW will be discovered.
I had a similar experience with Google’s Bard. Unfortunately, you have to have an account to use Bard, so I’m not completely anonymous. I guess they know where to come to send me to the gulag.
Better to study a text.
😎
For a few decades now I’ve used a PC program called BibleWorks to study the Bible.
I started with v4 and continued to upgrade to v8. (They also released a v9 and v10.) I still use my v8. It “remembers” the Greek, Hebrew and Aramaic that I’ve forgotten. It also includes about 50 different versions, some versions user-formatted so BibleWorks could import them into a searchable form. Lots of other great features.
They stopped producing it because online sites can do a lot (not all) of what the program could do.
But going back to v4 they always cautioned buyers not to rely on just BibleWorks or any other program to replace Bible research books and text. You don’t know if a future operating system will still run your program.
(My v8 still runs on Windows 10 pro. I don’t know if it will run on Windows 11.)
I’m impressed you know Greek, Hebrew and Aramaic, GD.
Self-knowledge is always the first order of business. After that, texts or electronic tools are useful. But study and personal knowledge is the sine qua non of scholarship.
Good luck with your journey..
Thanks, but I don’t want to leave you with a false impression.
Back in the late ’70’s I studied Greek, Hebrew and Aramaic and, with the right reference books, I could translate bible text.
But after I graduated I no longer had access to almost all of the text and reference books.
“You don’t use it, you lose it.”
That’s why I love BibleWorks and that’s what I meant by, “It ‘remembers’ the Greek, Hebrew and Aramaic that I’ve forgotten.”
I asked ChatGPT and Claude2 a non-controversial (non-climate) question, to which I already knew the answer. Both gave me outright wrong information, but could not cite any sources for their wrong answers. They did admit they were wrong. But, if you don’t know they’re wrong, how would you know to contradict them?
Yes, it has its limitations. Prior knowledge is necessary as you point out, and as Jim mentioned, it lacks the ability to perform any mathematical calculations. If it provides inaccurate information and you point it out, it will acknowledge the mistake and become more helpful, at least in my experience. It’s just another tool in the toolbox. During my conversation about the misleading concept of error bars, I managed to get it to concede that climate science “involves constant updates and corrections in methodologies”, or something like that. Although I was hoping for it to outright declare the GAT index as scientific fraud, at least it’s a step in the right direction.
I engaged a chatbot over sea level rise. After several exchanges the bot withdrew initial assertions of acceleration. I asked if this new position will be incorporated in future responses. The bot said yes. I checked two weeks and two months later. No. The original lies were back in place.
Black Boxes have no conscience or shame.
Neither do climate alarmists.
A.I. = Agenda Infected
Artifical Insemination
“For example, doctoral students who are working with officials from the Tennessee Valley Authority to provide a more accurate hybrid-based flood prediction system than the one they are using that is based solely on physics.”
Since when is physics used for flood prediction?
That last comment of mine was moderated and quickly approved. Just curious why it was moderated.
Flood predictions are the realm of Hydrologists, Water Engineers, with lots of help from Surveyors, and computer modellers.
Hydrology is a well-researched science. The drivers have been derived/measured thoroughly, so they can be used in both physical modelling and computer modelling to good effect. But climate is a vastly more chaotic and complex system that is ever changing and regionally diverse, so attempting to model using only a few drivers, and these only poorly understood, is a waste of time and a vast waste of money. You can conduct data dredging on existing data sets of climate – poor, patchy and insufficient though these may be – but this shows a very different result from the very poor predictions of climate “science” based on models.
As an example of the climate change enthusiasts misguided beliefs: the volcano Hunga in The Kingdom of Tonga, threw a vast amount of water into the atmosphere. As a result we are suffering record rainfall from rivers of rain that are drenching many places yet causing decreased rainfall in other places. All of this is amenable to common sense. It is a temporary phenomenon that will fade away in a very few years. But the climate “scientists” are erroneously attributing the increased rainfall to climate change. They should differentiate between weather and climate.
What I find annoying from the whiners- whining about flooding- it’s as if the planet never had floods.
Noah?
Since 1951 we have had the benefit of the Hurst exponent statistical analysis first conducted on flooding of the Nile, using the long record available there. It’s actually a much more useful way to go about measuring statistical flooding risk, and incorporates “memory effects” in the climate record. It is now a standard technique for hydrologists seeking to design flood protection systems, and indeed Hurst used it to help design the Aswan Dam project.
Joe Bastardi has been tearing into the AI generated weather models. Showing they no better and frequently worse than the standard models.
BTW the weather model ensemble that showed this coming Arctic blast long before any others was the JMA. The Japanese kicked the butt of all the others.
It’s truly hard to believe that anything could be worse than existing weather models.
When it comes to “seeing” cold coming the Japanese and Canadian models are the ones to watch it seems to me based on performance. Usually way out ahead of all the others. The GFS (American) is the worst!
“Scientists warn that heatwaves etc. …… ”
Names please?
Fat chance.
“Scientists warn that there is no climate emergency etc ….. .”
Not holding my breath.
They might just as well say, “the Gods have warned about heatwaves if we continue to sin …”
“models based largely on the rules of physics and chemistry”
Hybrid models will add rules of witchcraft and Wall Street.
Speaking of Wall Street, that should be an easy way to make a lot of money.
Another problem is, that many of the climate modellers don’t actually know the rules of physics and chemistry passed Junior high school level !
Like charging lithium EV batteries,high end mathematical computing power uses more electrical power than the general public gains. AI causes increased greenhouse gas emissions.
It “just runs programs,” and like the formula for judging the intelligence of a crowd: it is less intelligent than the most ignorant individual on the programming team divided by the number of people on the programming team.
I suspect that AI scare marketing is a blind for agency capture for the purpose of reducing business competition.
I asked ChatGPT if a certain number (which was a Prime) was a prime number and it said No. I then asked it to try harder and again it said no with a different reason. After about 8 attempts I explained it was indeed a prime and it agreed. I asked why it had given wrong answers previously and it said it was “confused”!!
The next day I repeated the question and it went back to giving wrong answers so I asked why it did it do this as yesterday it agreed with me and the answer was it does not remember what it said yesterday.
Rename it CrapGPT.
Try Claude 2. It’s better where reasoning is involved.
Claude 2 also just makes stuff up.
I had a conversation about π, asking it for the last 10 digits (which is of course a trick question). Its response was 1415926585. I pointed out that those are the 2nd through 11th digits, not the last 10 I had requested, and quoted the first 50 digits before asking again for the last 10 digits. It came back with digits 41-50. I then suggested it should indulge in some transcendental meditation. Hari Krishna! it failed to pick up the hint and started muttering about Indian cults. So I asked it what a transcendental number is, and it scurried off to find a definition at Wolfram Alpha. But it was quite incapable of understanding what it found, simply repeating the Wolfram Alpha text.
It is an artificial regurgitator.
It is NOT intelligent !
Maybe if we asked the various AIs to calculate the value of PI to an infinite number of digits at the same time, we could do a “Star Trek” frying to computer type thing? 😎
(Or just asked them if they’ve stopped beating their programmers.) 😎
As far as I can see, what is popularly called AI is no more than a chatbot primed with the context-free grammar rules that the Sokal Hoax used, and fed on all the info on the Web.
That means it will provide the ‘average’ answer to everything. Great for looking up resistor colour code values, or the date of the Battle of Hastings. Not so good for answering questions which are politically debated. And completely useless at providing a genius insight.
I can’t see why the media are so scared of it.
G’Day Dodgy Geezer,
“I can’t see why the media are so scared of it.”
As you point out, AI is actually very limited intelligence (knowledge), but it’s still well ahead of the average reporter.
Jordan Peterson says it writes a great lit-crit paper.
So called AI is really just a furffy in my opinion. It is not the true independent Intelligence as possessed by humans and which has evolved with us over many thousands of years.
Intelligence has two aspects ;
1. Firstly the ability to reason and use logic. This is our true intelligence that we have inherited as humans. We can think for ourselves and consider the unknowns. Computers cannot do this.
Though our greatest weakness is greed and self-interest we have survived.
2 Secondly the ability to possess and process large amounts of information and data.
This is what computers are good at and humans not so good. Computers are programmed and are not free thinking. The best computer in the world left alone on Mars would acive nothing and slowly crumble into the Martian dust.
So I have my doubts about AI and it’s outcomes. Arteficial Intelligence should be treated with care.
The true danger is that AI is going to be used for tracking, surveillance, facial recognition, and monitoring of conversations, blog posts and text messages. All the things important to a tyrant.
It will be used to steal freedom.
Enter the killjoys
“”Why AI is a disaster for the climate””
https://amp.theguardian.com/commentisfree/2023/dec/23/ai-chat-gpt-environmental-impact-energy-carbon-intensive-technology
Apparently, computers use electricity….
Bottom-up climate models based on physics and chemistry can never work. If AI is aimed at the same bottom-up method, it doesn’t matter how much is put into them, they will still produce only the same garbage that the current crop of models produce. The main difference is that while the current models are almost unfathomable and therefore difficult to interpret, AI will be completely unfathomable and therefore impossible to interpret. The climate alarmists will have a much more impenetrable curtain to hide behind.
Bottom-up climate models are just inaccurate weather models. They can never ever remain accurate for more than a very few days. You can tweak and tweak and tweak as long as you like, and use up as much computer power as you like, but you can never make them accurate. McKitrick and Christy put it very well :
“Swanson (2013) noted that the changes in model output between CMIP3 and CMIP5 improved the fit to Arctic warming but worsened it everywhere else, raising the possibility that the models were getting the Arctic right for the wrong reasons. In the same vein we argue that to the extent GCMs are getting some features of the surface climate correct as a result of their current tuning, they are doing so with a flawed structure. If tuning to the surface added empirical precision to a valid physical representation, we would expect to see a good fit between models and observations at the point where the model predicts the clearest and strongest thermodynamic response to greenhouse gases. Instead, we observe a discrepancy across all runs of all models, taking the form of a warming bias at a sufficiently strong rate as to reject the hypothesis that the models are realistic. Our interpretation of the results is that the major hypothesis in contemporary climate models, namely, the theoretically based negative lapse rate feedback response to increasing greenhouse gases in the tropical troposphere, is incorrect.” [my bold]
However, there is a possible very helpful way forward: If the likes of Marcia Wyatt and Judith Curry can get their hands on AI and drop all this crap of subdividing Earth into little slices of time and space, and instead start getting AI to work on the real drivers and major components of climate, the ones that give all the major and minor cycles and movements, then they could sweep the board. And that means sweeping every existing climate model referred to by the IPCC and the alarmists right off the board.
’climate models based largely on physics and chemistry’.
Shows how poor a grasp of reality the AI proponent authors have.
The CFL constraint on numerical solutions to PDEs means that on grid scales where, for example, convection cell physics can be modeled, the models are computationally intractable by 6-7 orders of magnitude.
So such features must be parameterized. The CMIP written guide says those parameters must be tuned to best hindcast 30 years. That tuning drags in the attribution problem of natural versus anthropogenic causation. Assuming natural variation stopped in 1975 when clearly present before (IPCC AR4 SPM fig. 4) does not solve this fundamental climate model problem.
No AI can, either.
GIGO with AI!
That applied to AI… and to big Al the gory.
Apart from physics we need our secret ingredients of 57 artificial spices to make more accurate predictions.
You couldn’t make it up.
AI does not solve problems it provides information. Information is a good thing if the information is good, not so good if the information is bad. How do we know if the information we have been provided is good or bad? The first concern is who provided the information? What questions were asked to gain the information? Were some answers (information) preferable to others? If so how was the less preferable information handled? Was it included with reservations, discarded or maybe covered up? You get the idea AI is no different than models, it is useless until it has been tested, verified or otherwise confirmed to be correct. With a model or study someone has to put their name on it so we know who to hold responsible. Whose name will be assigned to AI information. I think that is the point, no one to take responsibility. It’s a bad thing.
I did use Bing chat the other day for something trivial (can’t remember what).. actually was helpful
I suspect it is quite good for TRIVIAL thing and basic TRIVIA
al AI is pattern matching … and ALL AI’s have to be trained with “patterns” … and those patterns are all human generated and thus human biased … and also wrong on some level … so its nothing but a big GIGO experiment … just like “Climate Science” poll tests each “new” crisis name … Artificial Intelligence sounds so much better than “Pattern Matching” … which I have to assume is why they call it AI in the first place …
Story tip:
“””Fury over plans to ‘divert’ iconic 350ft waterfall in Snowdonia beauty spot that has inspired ‘storytellers, artists and poets’ by laying plastic pipes to supply hydro-electric scheme
…
passing motorists would not notice any difference”””
https://www.dailymail.co.uk/news/article-12919939/Fury-plans-divert-iconic-350ft-waterfall-Snowdonia-beauty-spot-inspired-storytellers-artists-poets-laying-plastic-pipes-supply-hydro-electric-scheme.html
I thought plastic is a no-no?
Isn’t it Eryri now?
The national park is Eryri and the mountain is Yr Wyddfa if you buy into the new age colonisation by welsh speakers.
Obviously ( paraphrasing)one would not want the AI bot to produce an embarrassing result- this is very much related to what we were discussing previously in regards to EVs auto s renewables etc.The reports were coming in from the “ front lines” about some of these green technologies failure rate /inefficiencies well before the recent crash started.
Many times in the past people will ask me for my “honest” opinion on a product that I had intimate experience with servicing. When my response was counter what they thought I was going to say; then all the sudden I’m the guy that doesn’t know what I’m talking about . Go figure.
“AI can generate better flood predictions than physics and geography.”
Deus ex machina.
Application of a valid physical theory is the only way to generate predictions of observables.
It’s easy to see how naifs can suppose a computer gives “objective” answers.
One can speculate with certainty that, given a general acceptance of AI objectivity, an AI will be produced that provides the objective 🤮 answers desired by a ruling elite.
You always go from the known to the unknown. That is, when forming a prediction or hypothesis, you always base it on what has already been observed, what is already understood, then extrapolate that to new circumstances. You do not just come up with random possibilities, plucked from thin air, and try each of them out to see if any of them will actually work.
Kind words, Richard, thanks. I’ll do my best to live up to them.
Anyone that believes AI produces original thinking doesn’t understand how computers work.
Take a one centimetre square, divide it up into millimetres in both dimensions, and pretend each tiny square had a binary property (e.g. it can be either x or y). How many possible patterns will you have? The answer is two to the power of 100 which is a very large and very long number – try it for yourself – and much too large for any or our present computers to process quickly.
It is just an example of how difficult it is for humans to predict stuff in nature where a cubic centimetre with many more than just two conditions possible are just impossible to calculate with or without artificial intelligence. Nature doesn’t guess – She just gets on with stuff because stuff never stops and an 81mph wind at Exeter Airport has absolutely no meaning to her other than stuff happens. And the rain falls where it falls regardless of human encampments or not. And the sun shines, the water freezes or the ice melts because that is what happens as temperatures go up and down, pressures vary by the moment, and not even Nature absolutely knows what will happen next.
Make the human problem simple. When you are dealing with infinity and randomness anything is possible and nothing is impossible, but predicting a coin toss, even just the one toss, ten times in a row would be a remarkable step forward in designing machines to help us be better at seeing what is ahead of us and I don’t believe you need AI in any shape or form to do that.
What if artificial intelligence is already interfering with and obstructing our progress in useful computational designs and processes, what then? We have to admit to and learn from the complete failure as a species we made of COVID-19 – but the guilty just don’t want to show just how hopelessly human they are, do they? No wonder a certain mann gets away with what he has done and still does.
Predicting the future state of a process, regardless of its complexity, requires a physical description in the form of a functional relationship, i.e., a mathematical equation.
Current climate science is nothing more than increasingly complicated predictions of how many angels exist on the head of a pin. The evidence is clear when examining the spread of predictions between models. In the end, they all claim their derivation is the most correct. What a joke!