Guest essay by Eric Worrall
A Climate activist explaining how AI technology will solve all the world’s problems. But the author kind of glosses over some of the limits of current generation AI technology.
How Artificial Intelligence Can Power Climate Change Strategy
Bernard Marr Contributor
Enterprise Tech
Jan 4, 2021,12:16am ESTSlowing down climate change is an urgent matter. If we fail, our world will face a more extensive crisis than we experienced because of the global COVID-19 pandemic. When artificial intelligence (AI) technology helps solve a problem, problem-solving can be done quicker, and the solution is often one that would have taken longer for humans to discover. Could artificial intelligence power climate change strategy? Yes, and it’s already doing so.
AI Can Accelerate Our Response to Climate Change
There’s no time to waste: atmospheric CO2 levels are the highest ever (even with significant drops from the stay-at-home orders for COVID-19), average sea levels are rising (3 inches in the last 25 years alone), and 2019 was the hottest year on record for the world’s oceans. …
Improve Energy Efficiency
According to the Capgemini Research Institute, artificial intelligence should improve power efficiency by 15% in the next three to five years. Machine learning supports efficiencies in power generation and distribution …
Optimize Clean Energy Development
In the Amazon basin, developers of hydropower dams have typically developed one at a time with no long-term strategy. …
Avoid Waste
Companies, governments, and leaders frequently deploy AI solutions to avoid waste. Whether AI is used to reduce energy waste from buildings …
Make Transportation More Efficient
Another quarter of global CO2 emissions is from the transportation sector. AI is already the technology that powers autonomous vehicles …
Tools to Help Understand Carbon Footprint
They say “knowledge is power,” and when it comes to climate change mitigation, AI can help build tools to help individuals and companies understand their carbon footprint …
…
Read more: https://www.forbes.com/sites/bernardmarr/2021/01/04/how-artificial-intelligence-can-power-climate-change-strategy/?sh=722ad9553482
Can Artificial Intelligence really do all this?
The following is a demonstration I created of something AI is good at, solving optimisation problems. In this case the AI is solving the “Travelling Salesman” problem, using an Evolutionary Algorithm.
If you imagine all the blue dots are cities, the AI rapidly attempts to work out the shortest route for a travelling salesman who has to visit all the cities, much faster than a human can – though AI also makes mistakes which a human can spot straight away, but which the AI struggles to identify and correct. The persistent loops which sometimes appear in the line are mistakes the AI failed to identify.
This is very similar to what happens inside your vehicle satnav, when you ask it to find a route to a destination. Satnav are a terrific aid to navigation – but we’ve all been in situations where the Satnav gave us directions which were plain wrong.
AI will help improve transport in the future, it can be used to unwrap and improve congested roads, or correct poor waterway planning, or inefficient building heating, or any number of other problems. All of these are optimisation problems, just like the travelling salesman problem – something AI is really good at – though a human would still need to review the AI solutions, to identify and reject solutions which contain mistakes.
AI is not going to solve the big problems in climate policy anytime soon, such as preventing blackouts with a grid supplied mostly by intermittent renewable sources. Such a solution, even if it is possible, would in my opinion require a level of creativity and comprehension of the issues which is well beyond the capabilities of current generation AIs.
As I’ve said before, AIs, for all their marvellous capabilities, are still currently just insect level intelligences, at best they have an insect level comprehension of the problem they are being asked to solve. Like termites building a mound, or ants building a nest, AIs can produce remarkable looking solutions to intricate problems.
But we all know what happens to insects when they encounter a problem which is beyond their comprehension – they splat into the wind shield. Or into confusing white surfaces, like the Tesla pictured in the accident at the top of this page.
If you were to start the AI program with correct information of where, how and why human life evolved on this Planet, then the only logical conclusion would be for AI to recommend a warmer climate over the entire Earth.
Fight Climate Fear. Warmer is Better.
Current AI is just Machine Learning. You give the algorithm lots of data items and a verdict on each and hope it reaches the correct verdict by itself on the next data item. It’s a lot of work to prepare application to common problems. There is no application to non-common problems.
There have been “traveling salesman” computer programs in service and commercially available for many decades by trucking firms and others. Call it AI if you will, but it they are just algorithms created by clever humans.
The Traveling Salesman program is NP-hard, meaning that it gets exponentially harder to find the best solution as you scale it up. Previous programs that dealt with it (and which are indeed human-designed algorithms) can only deal with finding the best solution when the problem is relatively simple, although they can often find good enough solutions with more complex problems.
I was peripherally involved in a project last year that used machine learning methods to try to do better. These ML methods did indeed work better for most cases, and it would be incorrect to characterize their working as “algorithms created by humans.” There are of course still cases that neither method works well for, although those tend to be rare(r), and of course AI doesn’t overcome the NP-hard problem in general. Neither do people.
“our world will face a more extensive crisis than we experienced because of the global COVID-19 pandemic”
It’s worse than we thought until Covid19 came along so now obviously it’s worse than that folks!
Super computers created the problem, so super computers will solve the problem. 😉
Virtually every electrical appliance is far more efficient than it was, say 40 years ago. Does the world produce less electricity than it did 40 years ago?
AI is good for chess…and ad targeting, and a lot of things, but it can’t change the law of supply and demand.
I can remember laughing decades ago about “Why do you need AI if you have the real thing ?”
“Artificial” Intelligence can solve Global Warming?
Didn’t “Artificial” Intelligence start this whole mess?
(Ever watch the original Star Trek? How often did they save a society but destroying/breaking the hold of the AL … er … AI that controlled them?)
Where is the definition for AI and how it differs from other computer programs with logic tests and inequalities?
Artificial intelligence does not exist and will not exist until a computer can have an original thought. Until then, it is just a machine with lots and lots of code containing lots and lots of IF conditions that the machine steps through mechanically without ‘thinking’ about what it is doing.