AI and biodiversity both thrive on complexity. For AI, massive amounts of data fuel machine learning models, helping them to recognise patterns, make predictions and solve problems. In most cases, the richer and more diverse the data, the smarter and more adaptive AI becomes.
Nature operates in a similar way. Biodiversity – the variety of life in an ecosystem – makes environments resilient and adaptable. Just as AI relies on diverse data inputs to function effectively, ecosystems rely on biodiversity to maintain balance, withstand disruptions and evolve over time.
Both AI and biodiversity highlight the power of diversity in creating systems that are strong, pliable and capable of solving complex challenges. As we head into the 16th meeting of the Conference of the Parties to the Convention on Biological Diversity (COP16), it’s clear that not just diversity of information but also the ability to process and understand it could be key to meeting the collective goals of protecting the planet’s biodiversity established by United Nations members.
The ‘Paris Agreement for Nature’
Last week, environmental leaders from nearly 200 nations made their way to Cali, Colombia for COP16. The event marks the first major UN gathering on biodiversity since COP15, hosted in Montreal, Canada in 2022, where participating countries established a landmark global framework for stopping the loss of natural habitats.
The agreement, signed by 195 nations and dubbed the “Paris Agreement for Nature,” requires signatories to seek to set aside at least 30% of their territories for conservation efforts, including additional protections for degraded ecosystems. Two years later, the goal of COP16 is to review the plans of countries that signed on to the agreement to see if the world is on track to meet the 2030 deadline.
Early indications suggest that ambition might surpass success. Signatories were expected to submit national biodiversity strategies and action plans (NBSAPs) before the start of the conference. Only 25 out of 195 countries filed plans to the UN Biodiversity Secretariat – and just five of 17 “megadiverse countries,” which are home to about 70% of the world’s biodiversity, have offered their pledges.
NBSAPs: A challenge of information overload
NBSAPs are a key factor both in understanding the habitats that nations have under their stewardship and in determining how to consider their protection into all aspects of governmental decision-making. Conservation is a cross-sectoral effort, and NBSAPs allow nations to take these conditions into consideration during policymaking.
They also present a unique challenge because there are so many conditions to consider and so much information to parse.
As Bernadette Fischler Hooper, head of international advocacy at WWF UK, points out: “We know change isn’t easy and countries are facing challenges such as a lack of funding, insufficient data and political instability.” However, experts say that technological solutions can help address some of the challenges nations face when trying to establish and follow through with NBSAPs.
AI’s role in processing data
In the two years since countries signed the original Kunming-Montreal Global Biodiversity Framework, there has been an explosion in innovation in artificial intelligence. As AI systems have gotten more powerful, generative AI has become an increasingly functional tool for everything from the processing and summarization of information to predictive analysis.
As nations seek new ways to measure biodiversity within their borders and process the massive amount of data needed to develop conservation plans, AI can help us gain more insight into our planet’s needs, and how we can address them.
Adam Thompson, global sustainable finance and ESG offering leader at IBM Consulting, has seen firsthand how AI can transform how countries handle, process and understand information.
“In the past, the problem statement was, ‘How to effectively work on arbitrary geospatial problems across heterogeneous datasets ranging from earth observation via weather model output to IoT sensors,’” he says. “And the volume of the data could be massive.”
Thompson says data from European Space Agency’s Sentinel-2 satellites – which collect new images of all of Earth’s land surfaces, coastal areas and inland waters every five days – amounts to about 3,2 Terabytes per day. Weather projections generated by the European Centre for Medium-Range Weather Forecasts could be as much as 250 Terabytes per day.
“From a capacity point of view, it is clearly is not feasible for a lot of organisations to manage such volumes of data for modeling purposes,” Thompson says.
Though it presents its own challenges when it comes to energy consumption, Generative AI (GenAI) offers a potential path for making data processing easier without stakeholders needing all of the data on hand.
Thompson says that one way to achieve that is through smaller models, like IBM Granite foundation models, which researchers can deploy for specific sustainability-related purposes, such as Earth observation and energy consumption.
“The accuracy of modeling using foundational models helps users improve results.”
A digital view into the natural world
In addition to processing data and generating predictions and analysis, AI tools can help us see how the natural world is changing around us. Earlier this year, IBM lent its IBM Maximo Visual Inspection (MVI) tools to WWF -Germany to track the movements of African forest elephants in the Congo Basin.
The African forest elephant is known as an ‘ecosystem engineer’, a foundational species in the overall health of the Congo Basin, which is home to more than 10 000 animal species.
The elephants are known for clearing out vegetation, making room for stronger and more resilient flora to thrive in an environment that would otherwise have its resources strained by overgrowth and competition.
They also help improve the diversity, density and abundance of plant and tree species, boosting the carbon storage capacity of the forest.
Similar technology has helped preserve coral reefs, which research suggests are under severe threat as the planet warms. The underwater structures, which an estimated 25% of all oceanic life count on at some point during their life cycles, help creatures in the ocean and on land.
Not only do they serve as massive carbon sinks, but they are also essential for preserving ecosystems that serve as food sources for humans. Over 4-billion people rely on fish that interact with coral reefs for 15% of their animal protein intake.
The Reef Company, which builds artificial reefs to restore reefs that have been lost to climate change, teamed with IBM to collect data on how the ocean is changing and where reefs are needed. Using IBM’s BluBoxx ocean data platform, the company can collect sensor data that measures a body of water’s salinity, temperature, pH, dissolved oxygen, pressure and carbon dioxide. That data can be accumulated and uploaded for rapid insights as well as deeper dives into how an ecosystem is changing.
AI and biodiversity: A perfect pair
We know more about our planet than ever before, thanks in part to AI-powered tools that allow us to gather and process massive amounts of information from multiple different sources – data that can be a challenge to manage all on its own.
“Today, we have massive amounts of data, but it’s not always accessible, relevant or consistently computed to scientific standards,” Thompson says. “We need trustworthy and transparent data for risk and opportunity identification and corrective actions.”
Luckily, much like a diverse ecosystem, AI can benefit from having access to more data. While the need to mitigate AI-related energy challenges remains clear, the technology does offer the potential for better understanding the world around us by processing the information we have access to and distilling it into a form we can understand and act on.
“Results need to be in a consumable format that the average person on the street can also understand and not require several PhDs to be able to interpret what the data represents,” Thompson says.
As nations set out to produce NBSAPs that will guide them in their goal to meet the standards set by the Kunming-Montreal Global Biodiversity Framework, AI tools may offer the opportunity to better understand what is happening in the natural world – and how we can conserve it.
Article by AJ Dellinger, tech reporter at IBM