Can AI solve climate change?
“It can only be attributable to human error” - Hal 9000, 2001 A Space Odyssey
TL;DR
Exploring the possibilities that AI unlocks in the fight against climate change
3 AI tools to make you and your team more efficient - Resquared, Forecast, UIzard
AI news: AI-powered... John Lennon? What happens when you make a copy of a copy (of a copy of a copy)? Google disagrees with Microsoft (I agree with Google)
Silicon vs carbon
Every day, every hour, every minute, someone somewhere is crying wolf about AI. A couple weeks ago we talked about how big tech is attempting to pull up the ladder - leveraging favorable regulation to further entrench their monopolies in the growing generative AI space. The primary mechanism they’ve taken to achieve this goal has been loudly blowing the whistle on abstract AI safety issues (think existential threats to humanity - Arnold Schwarzenegger in a leather jacket and sunglasses style), branding their tech as more advanced than it is while simultaneously increasing the public’s fear of AI (you can read that article here for more context).
But amidst all of the talk of synthetic sentience, human extinction, and mass unemployment, surprisingly little is said about what AI could actually help us accomplish. Beyond the profit margins, the branding and the buzzwords - what do we stand to gain from AI as a society?
When sitting down to think about how AI could possibly help us solve the major problems facing society today, I decided to narrow the scope of my research down to one existential threat in order to cut out some of the noise and avoid having a mental breakdown. I landed on climate change - as it’s a fundamental, undeniable, and largely apolitical (outside of the US at least) problem which all humans on the planet Earth should have some stake in seeing resolved.
I wanted to get to the bottom of this, so I decided to go to the most unbiased, rational, scientifically-minded place on all the internet... Reddit.
^the average Redditor in their natural habitat
In all seriousness though, I was able to find some unique arguments both for and against AI as a catalyst in our fight against climate change - many of which are actually supported by research being done by... you know, non-Redditors. The strongest arguments for are as follows:
Decreasing CO2 at the source
AI is potentially a gold mine for finding innovative new ways to reduce the production of CO2 at their source - for instance, in the high-polluting industries of steel & cement. It could also be instrumental in engineering the next generation of fuel-efficient gasoline cars as well as helping to speed up the development of electric cars and other more sustainable methods of transit and the infrastructure needed to bring them widespread adoption. You can read more about AI’s foray into the steel making industry in this article for the Carbon Herald: Carbon Re To Adapt Its AI Carbon Reduction Tech For Steel Making.
Making recycling more efficient
AI-driven computer vision and advanced robotics paired together could just be the miracle that the recycling industry desperately needs. Not only would this prevent a tremendous amount of otherwise reusable waste from entering landfills, it would enable the ever-increasing amount of net new products that capitalism demands with less raw resource input. To my positive surprise, this one isn’t just hypothetical. In Jennifer A. Kingson’s Axios article AI-powered robots could fix recycling's biggest problems, she details how these systems are already boosting the efficiency of MRFs (material recovery facilities) by outperforming humans at differentiating recyclables from non-recyclables and even going so far as to be able to identify which recyclables are too contaminated to be used (think a peanut butter jar with too much pb left in it to be recycled) - leading to less contaminated recycling streams and purer outputs.
Revitalizing the power grid
AI could be used to help regulate the power grid. How does this help us fight climate change? Well, keep in mind that all of those exciting new breakthroughs in renewables mean very little if they can’t be reliably used. By predicting energy usage based on robust AI-driven interpretations of usage data which is currently only used to determine billing for utilities providers, AI could predict the scale of demand for electricity before it spikes - preventing blackouts and making the grid perform more reliably. Furthermore, as Hayden Field and Grace Donnelly detail in their article Here’s how AI could transform the US power grid for Tech Brew, macro factors like weather patterns will be increasingly disruptive to an energy grid which will be more and more reliant on wind and solar energy without realtime data informing decisions around how much power to generate and how to avoid blackouts. AI is obviously a promising solution for interpreting this data, and with the massive amount of investment flowing into the US power grid, this one honestly seems inevitable.
Other ways AI could help fight climate change include improving carbon sequestration, using AI to better analyze and act on unwieldy climate data sets, and even detecting wildfires early so they can be stopped faster. On the other side of the aisle, I’ve seen a lot of arguments positing that the issue was never in potential solutions, but rather a lack of political will to do what’s necessary to stop the progression of climate change, which is a depressingly salient point. But, I’d be remiss if I didn’t mention the most direct argument against AI as a solution to climate change... and that’s the fact that training just one large language model emits more than 626,000 pounds of carbon dioxide (MIT, June 2019) which across all sectors has made it nearly as bad a source of pollution as the airline industry.
So the emissions are certainly bad, but to add a little perspective - Bitcoin alone is expected to emit nearly 62 megatons of carbon dioxide each year for as long as it exists, and aside from the argument that some applications of the blockchain could be useful in developing more effective carbon credit markets, it’s hard to argue that it’s a viable weapon in the arsenal we need to fight climate change.
Note that I don’t mean to pick on crypto here, there are a lot of current initiatives in the industry aimed at decreasing emissions and a lot of them have been effective. But the point stands that at least currently, blockchain technology presents little to no value in the fight against climate change, despite emitting an amount of carbon comparable to that of AI.
In the end, I suppose the truth is that while AI could be a very potent tool, it’s not without its trade offs, and despite the potential wins it can deliver, remains but one tool in the tool belt for fighting climate change. The reality is, it’s kind of always been on us not to screw this thing up...
3 AI-powered tools to turbocharge your product team
🏙️ Resquared
The product - Resquared is an AI-powered platform designed to make prospecting local business prospects more efficient. It offers several new AI features aimed at streamlining communication and outreach. Some of these include a communication center for tracking responses and following up, an improved emailing feature that speeds up the process of reaching out to prospects, a meeting scheduler tool for hands-off meeting scheduling, and a simplified interface for ease of use - it pulls your whole local business lead mining & outreach process into one, unified ecosystem.
The use case - this could be a game changer for sales and marketing teams that need to manage a large number of prospects. The AI-powered features allow for rapid communication with potential clients, efficient scheduling of meetings, and a more streamlined interface that can make the prospecting process more manageable and less time-consuming.
Listen to my interview with Tyler Carlson, co-founder @ Resquared here
📈 Forecast
The product - Forecast is a project management tool that employs AI to assist in task management. It considers various properties of your tasks such as the project, previously assigned person role, time of creation, deadline, etc. to suggest suitable roles, assignees, labels, and estimates. The system learns from the previous data used and becomes more accurate over time.
The use case - Forecast could be highly useful in complex project environments where assigning roles, estimating time for tasks, and tracking progress can be challenging. The AI-powered suggestions could help project managers allocate resources more effectively, keep track of project progress, and make informed decisions based on the estimated time for tasks.
🖥️ UIzard
The product - UIzard has made a ground-breaking move by launching the "world's easiest-to-use design and ideation tool" that's powered by AI. It makes the design process accessible to all, it doesn't even require previous design experience
The use case - envision UIzard as a go-to platform for bringing all your creative ideas to life, whether it's generating mockups from simple text prompts, scanning screenshots of apps or websites, or even using its drag and drop functionality to place UI components. Its real-time collaborative feature also make it the perfect tool for working alongside your team and obtaining immediate feedback from stakeholders. If it lives up to its promises, it should reduce the wait time for design resources, thus empowering you to take control and move projects forward at your pace. This means turning an idea into a clickable prototype becomes a much more straightforward, streamlined process.
Chronicles of the circuit circus
Paul McCartney says A.I. got John Lennon’s voice on ‘last Beatles record’ - by Jenni Reid for CNBC. The big pull quote:
““So when we came to make what will be the last Beatles record, it was a demo that John had that we worked on, and we just finished it up. It will be released this year,” McCartney said.
“We were able to take John’s voice and get it pure through this AI, so that then we could mix the record as you would normally do. It gives you some sort of leeway.” The BBC said it is expected to be a Lennon song from 1978 called “Now And Then,” which McCartney has in the past expressed a desire to “finish.” Lennon was murdered in 1980.”
The AI feedback loop: Researchers warn of ‘model collapse’ as AI trains on AI-generated content - by Carl Franzen for VentureBeat. The big pull quote:
“Over time, mistakes in generated data compound and ultimately force models that learn from generated data to misperceive reality even further,” wrote one of the paper’s leading authors, Ilia Shumailov, in an email to VentureBeat. “We were surprised to observe how quickly model collapse happens: Models can rapidly forget most of the original data from which they initially learned.””
Google challenges OpenAI’s calls for government A.I. czar - Hayden Field and Lauren Feiner for CNBC. The big pull quote:
“At the national level, we support a hub-and-spoke approach — with a central agency like the National Institute of Standards and Technology (NIST) informing sectoral regulators overseeing AI implementation — rather than a ‘Department of AI,’” Google wrote in its filing. “AI will present unique issues in financial services, health care, and other regulated industries and issue areas that will benefit from the expertise of regulators with experience in those sectors — which works better than a new regulatory agency promulgating and implementing upstream rules that are not adaptable to the diverse contexts in which AI is deployed.””
See ya!
And there you have it, another week of exploration into the vast universe of AI - from its potential role in the fight against climate change, to the latest breakthroughs in AI-powered product tools, and the ongoing AI saga in the news.
The key takeaway here is not to let fear and misconceptions steer our discourse on AI. While it's certainly a tool that could be misused or misunderstood, it also offers a tremendous potential for positive change in a ton of important fields - not least of which being climate change. It's a tool like any other - it's up to us how we wield it.
Thank you for taking the time to read this week's newsletter. As always, feel free to share this with anyone who you think might find it interesting or insightful, and I can’t wait to see you next week!