Responsible AI should be a top priority for impact investors. Climate investing offers lessons

In January, Anthropic CEO Dario Amodei published “The Adolescence of Technology,” arguing that AI is powerful enough to reshape society but not yet mature enough to manage the consequences. He called explicitly for investor scrutiny and external governance. When a leading frontier AI developer is publicly inviting structured engagement from the capital markets, that is not a thought experiment; it’s a call to action.

The most consequential technology in a generation is being built and deployed right now – but  most of the responsible investment community is only getting started in addressing AI as a governance, risk and impact priority.

This is not for lack of interest. Impact investors are sophisticated enough to see that AI will reshape entire business models and industries, create new opportunities for positive impact, disrupt labor markets, explode energy demand, shift the locus of economic power, and redefine how capital is allocated across industries and geographies. But the truth is that many investors are waiting for someone else to build out the responsible investing infrastructure for AI rather than dive headlong into the fray. This is a mistake.

I know this posture well because I watched it play out over decades as another mega issue moved from a niche concern to a mainstream consideration in the capital markets: climate change.

There are clear lessons to learn from the climate finance domain in mobilizing investor action on responsible AI. The parallels between AI and climate are striking—systemic risks, complex techno-economic transitions, political and policy crosscurrents. At the same time, significant differences make it hard to do a simple apples-to-apples translation from one domain to the other. 

But unlike climate’s slow burn, AI deployment is moving lightning fast. We have a narrow, closing window of opportunity for the impact investing community to engage and help set the terms of debate for the next chapter of this technology’s evolution.

We already know how to do this

Responsible investors can leverage learnings from climate to leapfrog into a more sophisticated posture on AI. From corporate engagement on fossil fuel production to the development of climate-related financial disclosures and sector-specific materiality maps, investors are familiar with the tools, frameworks and approaches that worked for climate. AI needs all of that. And it needs it on a compressed timeline, because new models ship in months, not decades.

But don’t repeat climate’s mistakes

A new white paper from my firm Great Circle Capital Advisors, published in partnership with the Omidyar Network and informed by dozens of interviews with the people who actually built the TCFD, SASB, Climate Action 100+, and GFANZ, identifies what worked and what didn’t in climate finance, and extracts lessons learned that are particularly relevant for responsible investors tackling issues related to AI.

The successes are well-known: government sponsorship through financial regulators lent legitimacy; simple messaging that framed climate as a business issue built political coalitions; staged implementation over a decade gave the market time to absorb new tools and concepts.

But our failures in mainstreaming climate risk and pushing for corporate climate governance matter more right now. The backlash to ESG didn’t emerge from nowhere. It was fueled in part by mission creep. Frameworks drifted beyond financial materiality into broader social and political territory. Meanwhile, a proliferation of overlapping initiatives confused, rather than clarified, the landscape for investors. The unraveling of GFANZ, the withdrawal of the SEC’s climate disclosure rule, and ongoing legal challenges to fiduciary duty have created headwinds that are still blowing, and not just for climate-focused investors.

For AI, the lesson is blunt: if responsible AI investing is not framed in ways that appeal to both the responsible investing community and the wider capital markets, it won’t succeed.  This means having strong arguments to address AI governance and risks rooted in financial materiality.  We need to focus on building an approach to investor action that is robust, grounded in evidence, and feasible for investors to take on in the current political environment. Here are a few simple ways to start:

Begin engaging companies and fund managers now on responsible AI practices.  Don’t wait for the “AI TCFD.” Climate finance needed decades of foundational work before the big frameworks could take hold. What you can do right now is invest in the building blocks—advance and promote “good enough” definitions for responsible AI use, leverage workable measurement approaches, and look to evidence-based risk assessments where possible—the unsexy infrastructure that makes future frameworks possible.  See examples from leading investors like Norges and Railpen, and from investor networks like Reframe Venture.

Build coalitions, but build them differently. The investor coalitions we built in climate were organized around a relatively concentrated set of fossil fuel companies and their supply chains. AI’s value chain—spanning developers, deployers, and users across every sector—is vastly more complex. As Omidyar Network’s Chris Jurgens put it on a recent ImpactAlpha call, “There’s a really robust responsible AI community and really robust responsible investing community, but they’re not talking to each other enough.” Bridging that gap is the single highest leverage move the impact investing community can make.

Engage through financial regulators, not technology ministries. The TCFD’s genius was its sponsorship by the Financial Stability Board. The SEC’s 2026 examination priorities have already elevated AI from “emerging fintech” to a core area of operational risk. That is the channel to work.

Invest in a responsible AI stack. Invest in “Good AI”, in addition to AI for Good. Omidyar Network’s investment in Anthropic alongside the Ford Foundation and Nathan Cummings Foundation wasn’t charity—it was a demonstration that impact capital can secure a seat at the table of frontier AI governance. Project Liberty, ReframeVenture, and ImpactVC launched an AI Due Diligence Tool for VCs in late 2025 to bring rigor to a space where due diligence frameworks barely exist. Mozilla Ventures has built a $35 million portfolio around trustworthy AI. These are the early movers. They need company.

And use the technology itself. Norway’s $2.2 trillion sovereign wealth fund is already deploying AI to screen its 7,200-company portfolio for ethical and reputational risks. Rather than adapting climate-era spreadsheets to a new domain, responsible AI governance should be built with AI tools from the start—enabling real-time monitoring, dynamic risk assessment, and scalable oversight that climate finance could only dream of.

The real question

Like climate-related risks and opportunities before it, AI is being embedded in corporate operations across virtually every sector, while agentic systems that can make autonomous decisions are deploying faster than governance models can address them. But unlike climate, there is no universally accepted metric—no equivalent of greenhouse gas emissions—around which to anchor responsible investor engagement.

This is not someone else’s problem. If you manage capital with any intention of aligning it with impact, AI governance is now squarely within your fiduciary responsibility. The infrastructure that will determine whether AI is deployed responsibly or recklessly is being built over the next four years. The investors who show up now will shape it. The ones who wait for the playbook will inherit whatever gets built without them.

Climate finance taught us that getting started—imperfectly, iteratively, with clear focus on materiality—beats getting it right on the first try. The same lesson applies to AI, only faster.

______________________________________________________________________________________

Dan Firger is the founder of Great Circle Capital Advisors, a climate finance and responsible technology advisory firm. Great Circle’s white paper, “From Carbon to Code: Applying lessons from a decade of climate finance practice to responsible investor engagement on Al,” was published in March 2026 in partnership with the Omidyar Network.