AI and ML in the fight against climate change
Like it or not, climate change is everybody’s business. Shimmers of good news in tech, however, offer some hope: More than 350 startups worldwide are already taking action by implementing eco-friendly technologies to reduce their carbon footprint, and artificial intelligence (though not inherently green) is considered key in this effort.
AI has already proven itself in finance, healthcare, and cybersecurity, but the combination of AI and its trusty sidekick ML with climate tech might be even more of a game changer given investment. The pipeline of climate tech startups is maturing quickly and, as The New York Times recently noted, AI funding has reached the point of a “no-holds-barred deal-making mania.”
Climate monitoring
The UN’s most recent report on climate change is unequivocal: It is a “survival guide for humanity.”
Doubling down on AI could be one of our best hopes. According to Boston Consulting Group (BCG), AI-related technologies could cut global greenhouse gas emissions (GHG) by a whopping 10 percent. This transformative potential is largely attributed to AI’s exceptional monitoring abilities, which include collecting and processing climate-related data, detecting changes in the environment, and understanding the far-reaching impacts of climate change.
A great example? Saildrone, a California-based startup that builds AI-powered, autonomous sailing drones for monitoring and analyzing oceanographic and atmospheric data — sea surface temperature, wave height, and CO2 levels — that are essential to our understanding of the marine environment. Most impressively, these drones can serve as our eyes inside Category 4 storms, giving us never-before-possible visibility into the physical processes of hurricanes and their potential coastal flooding impacts. (And yes, all with zero carbon footprint.)
Need to hire more AI engineers?
We’ve got experts standing by who can partner with you in two weeks or less.
AI and ML technology can also be employed in projects that remove carbon dioxide (CO2) from the earth's atmosphere. Pachama, for instance, uses machine learning models to measure and monitor CO2 stored in forests over time and identify carbon credit opportunities, helping organizations find forest restoration and conservation projects to invest in across the globe and track the effectiveness of their investments. Pachama raised $55M in Series B funding in 2022, and among its 800+ clients are leading companies like Salesforce, Microsoft, and Airbnb. (Heads up: If you’re into orangutans, check out Pachama’s work in Indonesia’s Central Kalimantan Peatlands, which are apparently capable of storing 20x more carbon than most forests.)
Other companies focus on creating emissions inventories that assess and monitor the level of GHG emissions within a specific jurisdiction or industry. ClimateAI, which recently raised $22M in Series B, utilizes AI algorithms and remote sensing technologies to estimate emissions from energy production, agriculture, and land-use change — a great example of AI at work.
Filtering and structuring large data sets
Filtering the signal from the noise in large climate data sets is yet another AI forte. It can automatically prioritize and categorize information, integrate diverse data sources, and preprocess data for analysis. By visualizing complex climate data and developing predictive models, artificial intelligence also makes it easier for scientists and policymakers to identify patterns and anomalies that might otherwise go unnoticed.
Consider Cervest, a leading provider of AI-powered climate intelligence that helps organizations effectively structure complex environmental data and derive insights related to climate risks (“EarthScan”). Then there’s Descartes Labs, which uses AI structuring techniques to process and interpret vast amounts of data related to climate change, land use, and agriculture — “geospatial intelligence.”
A small but growing number of companies use machine learning algorithms to provide valuable insights into supply chain data. A perfect example is our client Inspectorio, an Apple-backed SaaS startup that identifies sustainability-related patterns and anomalies across companies' supply chains using its Inspectorio Rise platform. Through continuous data analysis and AI-powered filtering, Inspectorio Rise empowers the likes of Target and Crocs to make data-driven decisions and optimize their decarbonization plans for maximum impact.
Climate modeling
Integrating AI-generated predictions into climate models enables data scientists and environmentalists to predict extreme weather events such as floods, hurricanes, and fires and model their potential impact on the climate and ecosystems.
FireScout, for example, relies on AI to operate a network of high-resolution cameras strategically positioned in fire-prone regions in the western United States and Australia. By continuously monitoring the camera feeds and utilizing cloud-based machine vision, the system can identify early signs of fire and rapidly alert authorities and firefighting teams.
Looking to take your business to the next level?
A scalable software solution built by our engineers could be the answer.
Another company making waves in the climate modeling domain is Floodbase. This New York-based startup combines remote sensing data and machine learning to generate predictive models for the consequences of flooding — and it’s partnering with FEMA and the Silicon Valley Innovation Program to provide a “national, near real-time flood intelligence system.”
The rise of AI has also brought innovative tools to agriculture and farming. In addition to forecasting weather changes, artificial intelligence now assists farmers in optimizing crop planting and preventing yield loss. Projects like Taranis use AI-powered image analysis to identify and monitor crop diseases, pests, and field conditions with “imagery so clear you can count the spots on a lady bug”; crop intelligence enables farmers to implement timely and targeted treatment strategies.
Optimization of complex systems
Complexity is no bar to AI, which is why it thrives in systems with multiple variables requiring simultaneous control. And numerous AI applications in the net-zero power sector serve as prime examples:
- Power supply and demand forecasting (Bidgely)
- Optimization of electric power distribution, e.g., enhancing cooling systems in large data centers (Submer)
- Management of renewable energy assets (AutoGrid)
- Optimization of energy consumption in residential buildings (BrainBox AI)
Another noteworthy example: Google-backed Eugenie focuses on simplifying emission management. Powered by AI and digital twins, the company provides manufacturers with real-time emission data sourced from satellites, enabling them to track carbon emissions and identify ways to reduce their carbon footprint at the industry level.
For the agri-food industry, AI and ML solutions are a game-changer when it comes to monitoring and optimizing growing conditions. In addition to Taranis, mentioned above, there’s Bowery Farming (a New York-based unicorn backed by Google, Natalie Portman, and Justin Timberlake, among other investors) which uses machine learning to create the ideal conditions for growing produce in automated indoor vertical farms.
AI and ML’s big climate challenge
AI and ML both serve as powerful allies in the ongoing battle against climate change. But perhaps the biggest paradox is that computationally intensive technologies like machine learning and big data themselves contribute to the global carbon footprint due to their voracious appetite for electricity.
The solution? Integrating these emerging technologies into the sustainable cloud infrastructure. As my colleague Thomas Morgenroth recently noted, using renewable energy sources and implementing energy-efficient data center designs and green software could mark a crucial turning point in the quest for environmentally responsible technologies.
There’s a very uncomfortable “but,” though. AI and ML aren’t silver-bullet solutions that can single-handedly reverse rising sea levels and halt species extinction. They can optimize specific processes, sure, but they cannot address the core challenges of a (really slow) global transition toward sustainable practices.
That responsibility is on us.
