224_Can AI Put Us on the Path to Sustainability?
Adventures in Sustainable Living Podcast
E224
Can AI Put Us on the Path to Sustainability?
New things, the unexpected, circumstances out of our control and any sort of change are all things than can make us a bit uneasy. Perhaps it is nothing more than fear of the unknown. We feel as if we will lose something or maybe we will lose our present lifestyle all together.
One thing is certain. Change is always scary. World politics, the climate, price of energy, price of food are all things that change constantly. Advances in technology are also constant. What was top of the line today is obsolete tomorrow. Now we face additional new technology called artificial intelligence. Is this good or bad?
So join me for E224, Can AI Put Us on the Path to Sustainability?
Welcome back everyone to the Adventures in Sustainable Living Podcast. This is your host Patrick and this is E224 Can AI Put Us on the Path to Sustainability.
The majority of the people I know do not deal with change very well. I know this because of the type of work I do. I am often the person that has to implement changes. I am often challenging the status quo. Even though I am attempting to make changes for the better I still meet up with a considerable amount of resistance. This is because most people just don’t like change.
Sometimes I think I am the only one that does. But what most people do not know, or do not want to accept, is that change is inevitable. The use of AI is a perfect example. This technology will change so many things, perhaps for the better, but already people are afraid of it.
What I want to accomplish in this episode is to first look at the concept of AI and then discuss some ways in which this developing technology may help us to build a more sustainable future.
Good News Story of the Week
But before we get started with that, let’s talk about the good news story of the week.
It is likely no big surprise that the developed world continues to fail at confronting the biggest challenges of our modern age, one of which is climate change. But the good news is that the private sector is stepping in to pick up the slack to the tune of $100 billion.
The United Nations Environment Program reports that financing to protect nature is surging with an eleven fold increase in just four years. This has occurred in a variety of ways including managed accounts, electronically traded funds, debt-for-nature swaps and venture capital funding. If capital continues to flow at such rates, it is projected that by 2030 the private sector would have contributed over $1 trillion into investments that protect nature in some way.
While the governments of the world continue to fail to face up to and meet the biggest challenges of our modern world, it is nice to know that the private sector is working to pick up some of the slack.
So, let’s move on to this weeks episode regarding AI and sustainability. But let’s start out by looking at a tiny bit of history because I think it helps to bring things into perspective.
The First Computer
The first modern computer was built in 1942 by John Vincent Atanasoff and his graduate student Clifford Berry at Iowa State College, now known as Iowa State University. However the first programable, electronic digital computer, known as the Electronic Numerical Integrator and Computer, ENIAC, was completed in 1945.
This computer contained over 17,000 vacuum tubes, 70,000 resistors, 10,000 capacitors, 6,000 switches and 1,500 relays. I measured 8 feet tall, 3 feet deep, and 80 feet long and required its own air conditioning unit to manage all the heat generated during operation. And now it is amazing to think we all walk around with a hand-held computer that fits in our pockets, also known as our smart phones.
The Birth of Artificial Intelligence
In 1950, Alan Turing, an English mathematician and computer science pioneer, posed the question, “Can machines think?” To test this question, Turing devised a game which later became known as the Turing Test. It was an adaptation of a Victorian-style game that involved the seclusion of a man and a woman from an interrogator, who must guess which is which. In the Turing Test, one of the participants was replaced by the computer program. The interrogator had to tell which was the human and which was the computer. If the interrogator was unable to tell the difference, the computer was considering to be “thinking” or to possess “artificial intelligence.” It is interesting to note that Turing’s test came merely years after the development of digital computers.
In 1956, John McCarthy, a professor at Dartmouth College, organized a workshop to develop ideas about thinking machines. For this project he chose the name “artificial intelligence.” This Dartmouth workshop is widely considered to be the founding moment of AI as a field of research which focuses on “how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans and improve themselves.”
In 1961 an MIT alumnus James Slagle, developed a program called SAINT, symbolic automatic interrogator. The program is acknowledged to be one of the first computer systems that could emulate the decision-making ability of a human expert. Today, AI is referred to as a branch of computer science that aims to build machines, capable of performing tasks that typically require human intelligence.
In today’s world AI enables machines to simulate human abilities such as problem-solving, learning, decision-making and comprehension. Common applications include image and speech recognition, content generation and self-driving vehicles.
Human Fears over AI
As interesting as this all sounds there are actually a lot of human fears over the rapid development and application of AI. People fear that AI will surpass human intelligence and make decisions that are detrimental to human existence and potentially lead to catastrophic outcomes. There is the fear that AI will become too powerful and take societal control out of the hands of humans. People also fear that AI will replace human jobs, result in unemployment and disrupt our economy. However, most AI proponents say that these fears are often based on a lack of understanding and fear of the unknown. It is better to educate yourself on the potential impacts of AI and appreciate the benefits.
Sustainability and AI
Now that we have covered the basic concepts of AI as well as some of the fears and reservations over this technology, let’s talk a little bit about the potential benefits as related to sustainability.
AI can actually help people be more sustainable in their everyday lives in multiple ways. For example:
1. Smart Home Energy Management
AI-powered thermostats (like Nest or Ecobee) learn user habits and optimize heating/cooling to reduce energy waste.
2. Smart Lighting & Appliances
AI-driven smart plugs and lighting systems automatically turn off unused devices, reducing electricity consumption.
3. AI-Powered Route Optimization
Navigation apps (like Google Maps and Waze) use AI to find the most fuel-efficient routes, reducing carbon emissions from travel.
4. Personalized Carbon Footprint Tracking
Apps like Earth Hero or Oroeco use AI to analyze spending habits and lifestyle choices to help individuals track and reduce their carbon footprint.
5. AI in Sustainable Shopping
Retail platforms use AI to recommend eco-friendly products, sustainable brands, and second-hand options based on user preferences.
6. AI for Smart Recycling
Apps like Recycle Coach or Lasso Loop use AI image recognition to help users sort waste properly, reducing landfill contributions.
7. AI in Food Waste Reduction
Apps like Too Good To Go and Olio use AI to connect users with surplus food from restaurants and grocery stores, preventing food waste.
8. AI-Assisted Water Conservation
Smart irrigation systems (like Rachio) use AI to analyze weather and soil conditions to optimize water use in gardens and homes.
9. AI for Sustainable Transportation Choices
AI-driven mobility apps suggest eco-friendly transportation options like biking, walking, carpooling, or electric vehicle (EV) routes.
10. AI-Optimized Diet Choices
AI-powered apps (Yuka or GreenChoice) help users choose sustainable and ethical food options by analyzing product ingredients and environmental impact.
AI Assistance in Addressing Environmental Concerns
Additionally, AI can help to address environmental concerns in several ways. This technology can predict weather patterns and track climate disasters, enabling better preparation and response strategies.
AI can help improve agriculture by enhancing sustainability and reducing environmental impact. It can potentially boost farm efficiency by 20% to 40% through better weather prediction and more precise application of water, fertilizers, and pesticides, which can reduce the environmental impact of agriculture
AI can assist in monitoring global climate change impacts and progress on sustainability targets through satellite monitoring. This technology can also support environmental compliance.
AI can be used to track and manage waste more efficiently, reducing methane emissions from landfills. Greyparrot, a software startup, has developed an AI system that analyzes waste processing and recycling facilities to help them recover and recycle more waste material.
AI also contributes to environmental monitoring initiatives, such as the GEMS Air Pollution Monitoring platform, which aggregates data from over 25,000 air quality monitoring stations in more than 140 countries and leverages AI to offer insights on the impact of real-time air quality on populations
Furthermore, AI can be used to calculate the environmental and climate footprints of products, providing valuable data. Initiatives such as the International Methane Emissions Observatory (IMEO) use AI to monitor and mitigate methane emissions, potentially contributing to systemic change at an unprecedented speed and scale.
As far as energy management is concerned, AI can optimize building design and controls, potentially reducing energy consumption by 29% or more.
Additionally, AI can be a powerful tool in improving traffic flows, reducing idling, and dramatically reduce the number of accidents, contributing to sustainable transportation. AI also plays a crucial role in biodiversity conservation, such as monitoring coral reefs and providing real-time data analytics for marine biodiversity projects.
Obviously there is a wide range of applications for AI. But it is worth noting that these applications of AI highlight its potential to support environmental sustainability and fight climate change. However, they also underscore the importance of responsible and ethical practices to mitigate the technology's own environmental impacts.
The Environmental Impact of the Internet
While all of the potential benefits to AI sound appealing, this technology is not completely benign from a sustainability perspective. If you think that sitting at home on the internet has no impact whatsoever, you are actually quite wrong. That said, the environmental impact of the internet, in particular AI, was truly a learning issue for me as well. Once I started investigating this, I was quite surprised at what I discovered.
As a general rule, the primary environmental impact of the internet is its significant contribution to global carbon emissions. The internet and its supporting infrastructure account for around 4% of global carbon emissions, which is comparable to the aviation industry. If the Internet were a country, it would be the world's fourth-largest polluter.
This substantial environmental impact is largely due to the energy-intensive nature of the internet's infrastructure, including data centers, transmission networks, and the devices we use to access the internet. Data centers, which could be considered the factories of the information age, require a tremendous amount of electricity to operate 24/7, and many of these centers still rely on fossil fuels rather than renewable energy sources.
The use of digital technologies, including online video, social media, and cloud storage, further exacerbates this issue. A perfect example is online video which alone generates 60% of world data flows and results in over 300 million tones of CO2 emissions per year.
Efforts are being made to mitigate these impacts through transitioning to renewable energy sources and improving the energy efficiency of data centers and devices. Major tech companies such as Google, Facebook, and Apple have all committed to powering their data centers with 100% renewable energy, although many other providers have not yet made the same commitment.
In summary, the primary environmental impact of the internet is its substantial carbon footprint, largely driven by the energy consumption of its infrastructure and the increasing demand for digital services.
Certainly AI has been hailed as a game-changer. But underneath all the potential benefits that we focus on lies a different pressing concern. AI technology comes with a staggering carbon footprint associated with the development, maintenance and disposal of associated technology.
According to OpenAI researchers, the amount of computing power needed to train cutting-edge AI models has doubled every 3.4 months since 2012. This is primarily due to datasets and models becoming more complex. It is anticipated that by 2040 emissions from the Information and Communications Technology (ICT) industry as a whole will reach 14% of global emissions. Similar to the internet as a whole those emissions are directly related to data centers and communication networks.
In June 2019, MIT Technology Review published an article focusing on the carbon footprint of deep learning and comparing that to the oil industry. Researchers at the University of Massachusetts, Amherst, performed a detailed life cycle assessment for training several common large AI models. The paper specifically examines model training for natural-language processing (NLP). This subset of AI focuses on teaching machines to handle human language such as translation and sentence completion. OpenAI’s GPT model is just one prime example.
What the research revealed is that this whole model training process can emit the equivalent of more than 626,000 pounds of CO2, which is nearly 5 times the lifetime emissions of the average American car including the manufacture of the car itself or the equivalent of 300 round-trip flights between New York and San Francisco. But it doesn’t stop there.
An AI model is never trained only once. The first round of training is more or less considered baseline function. The fine tuning process is known as neural architecture search, which optimizes a model by incrementally tweaking a neural network’s design through exhaustive trial and error. The fine tweaking requires many more rounds of training and tuning all of which substantially increase the carbon footprint of any one AI model.
The next impact to consider is disposal of all the associated electronic waste, known as e-waste. This type of waste contains hazardous chemicals including lead, mercury, and cadmium, all of which pose threats to the environment as well as human health.
There are even concerns about the rising use of AI in agriculture and environmental management. Implementing AI in agriculture tends to focus on increasing yields but this could be at the expense of maintaining ecosystem health and further exacerbating biodiversity loss. For environmental management, decisions made by AI systems could be biased if they were presented with incomplete or inaccurate data. If an AI system was given instructions to value economic growth over environmental protection, it might choose short-term financial gain over environmental sustainability. And this of course is directly in-line with what humans tend to do anyway.
Final Thoughts
If you look back over the time period during which humans have been present on this planet, there are distinct turning points that literally changed our fate and history. The implementation of agricultural practices and the Industrial Revolution are just two examples. Maybe the implementation of advance AI is another turning point.
One thing is certain about humanity. We rarely seem to consider the long-term impact of anything we do. We are much better at a fight-or-flight response than we are at long-term planning. Since we are more or less at the early stages of advanced AI technology, perhaps now is the time to consider the repercussions of our choices instead of being distracted by the alluring advantages.
To reduce the impact of AI systems, first and foremost we must address the massive energy consumption and carbon emissions associated with developing, training and using AI models. This can be done by optimizing hardware and algorithms, promoting ethical AI design standards, and taking into account the impact of waste disposal. AI technology industries and governments must also adopt a culture of openness and responsibility to ensure the ethical creation, use, and disposal of AI technologies.
Despite the many human fears over the use of AI technology, there are many significant benefits. However, similar to any other significant advancement, we must choose carefully. By adopting a culture of openness and responsibility, we can easily promote human welfare and use AI to help us reverse and repair much of the damage to our planet posed by human activity.
Admittedly, I’ve had some reservation about the rapid development of AI. But at the same time, I have also used it to fine tune some of my own activities even on the homestead, on my blog and on the podcast. It truly is an incredible tool. If used wisely, we can all benefit from the incredible power of this technology and use it to help us build a sustainable future. But the environmental concerns of this technology are cause for genuine concern and do demand our attention. While the allure of this technology is undeniable, we must face up to and mitigate the negative impacts as well.
But like anything else we do, our choices make a significant difference. Once again, I think humanity is at a major crossroads. Given the right choices, we can use this technology to not only reverse climate change but build a future that is sustainable for many generations to come. The transformative power of this technology is truly alluring but we cannot ignore the environmental impact.
Well folks that’s about it for this week. I hope you have enjoyed this episode and will join me again next week when I am going to talk about how to reduce your personal environmental impact of using the internet.
Until then this is your host Patrick signing off. Always remember to live sustainably because this is how we build a better future.
The Green Dilemma: Can AI Fulfil Its Potential Without Harming the Environment?