AI and global climate policy often seem at odds. The crisis that climate policy is trying to address is driven by greenhouse gases, a problem potentially made worse by the sheer amount of energy and resources demanded by the explosive growth of AI infrastructure. But Michal Nachmany is convinced AI can be deployed for the benefit of society, business and earth’s climate.
Nachmany, who holds a Ph.D. from the London School of Economics, is the Founder and CEO of Climate Policy Radar (CPR), which has built the world’s largest open database of climate laws and policies covering all countries. CPR uses AI to augment its database, and to serve its more than 300,000 users who often work in parliaments, government agencies, the United Nations and NGOs, as well as financial markets and multinational companies.
Nachmany has always been interested in designing systems for the public good. Completing a degree in law and an MBA specializing in finance, she then earned another graduate degree in energy and environmental management. While working on a project to reduce greenhouse gas emissions, she found herself needing to know which laws and policies other countries were developing to manage this crisis. “I decided to write a Ph.D. about the dynamics of the diffusion of such policies,” she says.
Coincidentally, leaders at the 2009 Copenhagen Climate Summit COP15 had run into that same problem: an acute lack of information on existing climate policies. And they had asked the London School of Economics (LSE)—where Nachmany was studying for her doctorate—to collect and compare climate laws of the world’s governments.
“First, I stepped into that as a research assistant at LSE’s Grantham Research Institute, and then I managed that project a year later,” Nachmany says. During and after completing her degree, she managed this growing database, called Climate Change Laws of the World, out of LSE’s Grantham Research Institute. But to make the most of that data, she realized she needed better technology.
“When I go online, I can find vegan ice cream shops near me that are open right now,” she recalls thinking. “But I can’t find policy incentives for renewable energy that are in force in Southeast Asia. That makes no sense. Why not apply these same technologies to the world’s climate laws?”
That led her to found Climate Policy Radar in 2021. The first version of their search tool was released the following year. It has now grown to include more than 11,000 climate laws, policies, climate finance projects and international climate commitments from around the world, with plans to integrate ever more relevant documents, including corporate climate transition plans.
We spoke with Nachmany about her background, the future of CPR and the role she sees open algorithms playing in the world’s climate conversation. “We’re in the business of helping everyone get the data they need to make better and more informed decisions,” she says.
“While good information doesn’t necessarily lead to good decisions, bad or absent information, almost by definition, leads to bad decisions.”
The primary goals of CPR are around providing access to better climate information. Is that right?
Yes, with a focus on the pathways to action—the policies, the corporate transitions, the sustainable investment enablers. And while “climate” is in our name, we now always find ourselves adding “and nature.” Health, human rights, peace and security are all interconnected with policies on climate and nature. Policymakers are looking for information on how to better include these different dimensions in their policy development.
While good information doesn’t necessarily lead to good decisions, bad or absent information, almost by definition, leads to bad decisions. This is why we focus on enabling the two core levers that are needed in addressing climate change: better policies and moving money from the old carbon-intensive economy to the new low-carbon and zero-carbon economy. Our users span the entire ecosystem—from governments and civil society to researchers, journalists, corporations and financial institutions—because everyone is either making, challenging or being impacted by policy.
One of the things I’m proudest of is our work to unlock more climate finance for developing countries. We recently partnered with all four of the major multilateral Climate Funds, which are tasked with funneling climate finance from the rich countries to the developing countries. For developing countries, the process of applying for these funds is extremely complex and often results in long delays, or in the poorest countries not being able to apply at all. Government agencies in developing countries often aren’t sure which fund they should even apply for and how.
To solve this problem, we partnered with the funds to create a single clearing house for all of the projects that they finance and the detailed guidance for application. We did this by creating a single source of machine-readable, searchable documents. This will allow funds to be deployed faster, build clean energy projects sooner, and be more prepared for the growing frequency and intensity of extreme weather events. That is the impact we seek and AI plays a central role in that.
How does your platform serve business and corporates?
We can look at this through a lens of different types of risk. There are three types of risks related to climate and nature: physical risks, transition risks and the risks of liability.
Let’s look at physical risk: Assume you are a beer producer and need to assess the impact of droughts on hops harvest. How do you plan for this? Of course, there are climate models for predicting droughts, but it is also important to take the policies into account that can mitigate or exacerbate supply chain risks. These can be water allocation laws or subsidies for water-efficient crops that determine the availability and price of your raw materials. Or let’s assume you are a manufacturing company: Extreme weather events can disrupt your energy supply. To fully calculate that risk, you also need to factor in the policies that support your regional power grid’s resilience, incentivize your onsite renewables or battery capacity, or exacerbate your company’s reliance on price-volatile fossil fuels.
The interactive map at Climate Change Laws of the World, a site operated by Grantham Research Institute at London School of Economics, is powered by Climate Policy Radar’s technology and database. It offers the ability to search the data for information on countries’ climate-related policy documents, allowing decision makers around the world access to the best information.
Then there are transition risks. A manufacturer of auto parts or fast fashion must ask: Will my products still be marketable? If production of new petrol and diesel cars are banned in Europe after 2035, what happens to a business built around supplying their components? If the climate-related regulation of fast fashion is changing, how will that affect my business model? Being able to quickly map this shifting policy landscape is essential to understanding the business’s long-term viability and strategy.
The third risk is liability—the legal and reputational challenges that come with evolving regulations. Could a company or its investors face increased scrutiny for greenwashing, certain supply chain practices or its environmental impact? As regulatory frameworks tighten, understanding disclosure requirements and compliance obligations is essential to mitigating risk and maintaining trust.
And finally, there is the opportunity for companies to engage with policy design, which requires staying on top of regulatory developments. An emerging criterion of credible climate transition plans is how engaged the disclosing company is in also advocating for the public policies that support the decarbonization of their suppliers and clients. To develop such a corporate climate policy agenda, companies need timely and easily accessible information about emerging policies in their key markets.
In the coming weeks, Climate Policy Radar will expand its open datasets to include climate-related disclosures from hundreds of major companies. Enabling easy access to corporate disclosures together with data on laws and policies, national commitments, climate litigation and project financing, provides a clearer picture of emerging risks and opportunities—helping businesses, investors and policymakers make better decisions.
Does the growth of misinformation and disinformation have any impact on your work?
Absolutely—but I would add a third and fourth issue to those two: bias and omission. Together, these make up what are called information disorders, and they distort decision making at every level.
We aggregate public documents—laws, policies, regulations, corporate reports and litigation cases—and take great care in authenticating them. Bias, usually a result of training AI models on limited or unrepresentative datasets, as well as algorithmic omission, are major issues, especially in climate. If an algorithm doesn’t give you the full picture, key policies or legal rulings can be missing from the analysis—and you don’t know what you don’t know. And that’s not just an academic problem, it affects how we vote—through the media we consume—how regulations are shaped and how businesses plan for the future.
When it comes to our own use of AI, it is important to mention that Climate Policy Radar’s data and models are open for everyone to examine. That’s like cooking in front of the guests in a restaurant. They see what you put in the food and how you made it. In our Generative AI features, which will be available soon, users can trace every sentence in the generated answer to the source on which it was based. By exposing our inner workings this way, we are able to build trust, reduce bias and reduce misinformation, disinformation and omission.
It all comes back to this notion of a public good—in our case the transparency in our data and algorithms—and speaks to the importance of guarding the truth in any political system.
“By exposing our inner workings, we are able to build trust, reduce bias and reduce misinformation, disinformation and omission.”
Your website mentions the importance to your strategy of “radical collaboration.” Can you talk about that?
Radical collaboration means rethinking how we work as an industry. We see other organizations as allies, not rivals. Climate progress is too urgent for duplicate efforts, secrecy or inefficiency.
We focus on open science, shared standards and interoperability, ensuring that our work and our learnings can be used by others to drive impact faster. That includes running a community of practice on natural language processing for climate, organizing conference workshops and being good allies to other organizations in the space.
No single organization can do it all—nor should they. Collaboration works best when each of us plays our part in the value chain. Each organization does what they’re great at so others don’t have to. Some collect data, some build tools, some model and analyze, and others drive advocacy and policy change. We provide high-quality, structured climate data, so others can generate insights and take action. It’s an ecosystem approach of interdependence.
How are you funded?
Most of our income comes from philanthropy. Some of our biggest contributors are the Quadrature Climate Foundation, the Sequoia Climate Foundation, Google.org and the Patrick J. McGovern Foundation.
Funding tech nonprofits is still new to many foundations, as it is so different from classic NGO funding. We have been fortunate to work with our supporters on shaping that practice, for the benefit not just of ourselves but also of others.
In the future, we’ll offer paid API access for those who need scalable, high-volume integration, while ensuring the underlying data remains open and freely accessible. This model supports our financial sustainability without compromising our commitment to open knowledge.
How do you see CPR growing?
First, by expanding our data coverage. For example, we already have national policies but haven’t yet tackled state or city-level policies.
Second, by enhancing our analytical tools. We’re improving structured data analysis and advancing our Generative AI capabilities—from prototype to robust tools for question-answering, synthesis and summarization, all tailored for the climate domain.
Third, by evolving the platform for different types of users, from policymakers needing quick comparisons to researchers analyzing long-term trends.
Fourth, by building new alliances to include other types of data. Climate Policy Radar specializes in text analysis, but decision makers need more. They also need Earth observation data and supply chain data. Instead of sending them elsewhere, we’re leading a multimodal decision-support collaboration, bringing these data sources together. Expect more on this by London Climate Action Week in June.
Fifth, by continuing to foster responsible AI. We don’t just apply AI—we shape best practices for its responsible use in public good applications. This means improving accuracy, governance and oversight mechanisms while sharing our learnings with the wider community. Ensuring that AI serves climate action in a way that is transparent, ethical and impactful is core to our mission.
And, of course, the most important growth target we are working toward is to see ever more organizations, governments, members of parliaments, NGOs, corporations, UN agencies and scientists worldwide use our tools to accelerate climate action.