Sustainable investments are on the rise whether the industry is ready or not, writes Lorenzo Saa, Chief Sustainability Officer at Clarity AI.
As capital takes on sustainability, investors face growing pressure to leverage on high quality and reliable sustainability data, while rethinking how investment decisions are made. That means leaning into efficiencies, embracing innovation, and exploring smarter ways to drive performance. Artificial intelligence (AI) is already playing a central role in this shift. With its ability to unlock insights from complexity, AI can help institutional investors navigate an increasingly crowded sustainability landscape, identify risks and opportunities, and meet growing demands—from clients, regulators, stakeholders, and society at large.
AI is rapidly transforming how we approach sustainability. It’s no surprise that forward-thinking companies are already embedding AI deep into their operations—and investors are no exception.
As global sustainability investment grows, investors face a fragmented and often inconsistent ecosystem. AI has the potential to enhance investment decisions across the portfolio lifecycle: from sourcing and validating ESG data, to spotting emerging risks and opportunities, assessing impact, deepening research and automating reporting. But unlocking these benefits requires thoughtful risk management, strong governance, and a commitment to responsible innovation.
Enhancing data coverage and quality
Managing diversified global portfolios generates an overwhelming volume of ESG data. The challenge lies not just in collection, but in standardisation, verification, and usability.
AI and machine learning are already reshaping how data is collected and structured, enabling investors to draw from a wide variety of sources and formats—text, tables, graphs, even satellite imagery—and then stitch it all into something usable.
But quality control remains critical. Clarity AI’s research found that across three leading data providers offering the same reported emissions figures, results differed 13 per cent of the time—with discrepancies of more than 20 per cent in some cases. This is where machine learning-based reliability models become essential—flagging anomalies and validating sources to ensure data integrity.
What about when proprietary corporate data is missing? AI can step in—using alternative sources like news reports or geospatial imagery to fill gaps, offer new angles, or challenge what’s been reported. For instance, AI tools can cross-reference corporate human rights disclosures with independent news sources to uncover inconsistencies and strengthen due diligence.
AI can also produce estimations where no direct data exists. For example, Scope 3 emissions—which are notoriously underreported—can be inferred using AI-driven estimation models that go beyond sectoral and country benchmarks, allowing for more accurate, company specific modelling.
Turning data into insight
AI enables investors to go beyond passive data collection, offering interactive tools that enhance engagement with their portfolios. New AI-driven assistants allow asset managers to explore sustainability performance, interrogate data sources, and receive tailored recommendations for improvement—all in real time.
AI isn’t just backward-looking. It enables forward-looking analysis too. When we applied machine learning to examine the transition plans of the world’s 400 largest carbon emitters, only 40 per cent were found to have credible decarbonisation strategies—a critical datapoint for investors facing rising climate risk and regulatory scrutiny.
What’s more, AI is already helping investors report to multiple stakeholders in different formats and languages. Most importantly, it can streamline compliance processes—freeing up investors from repetitive tasks so they can focus on what truly matters: making smarter sustainable investment decisions.
Guardrails for innovation
The benefits of AI are significant—but there are risks. Like a high-speed vehicle, it requires steering and brakes.
To harness AI responsibly, institutional investors must establish strong governance structures to oversee its use. This means ensuring accountability, avoiding opaque decision-making, and putting human oversight at the core.
Disinformation and “hallucinations” (AI-generated falsehoods) are real risks. Systems must be built with safeguards to prevent misleading outputs. Privacy and security must also be top priorities—ensuring data isn’t accessed or misused without proper authorization.
And fairness can’t be an afterthought. AI models must be trained on diverse, representative data to avoid reinforcing bias. Without this, we risk entrenching inequality rather than solving it.
AI’s own footprint
While AI can accelerate progress on climate and nature-related goals, it also poses environmental challenges that must be managed—particularly its consumption of electricity and water.
Investors have a vital role to play in making sure AI’s infrastructure aligns with sustainability goals. That means advocating for workloads powered by renewables. Encouragingly, we’re seeing big tech companies—some of which have seen emissions rise due to AI usage—begin to respond. Microsoft’s partnership with Brookfield Renewable, along with Google and Amazon’s investments in small modular nuclear reactors, are promising steps.
Efficiency in model design also matters. Not every task requires a massive, resource-intensive generative AI model. In many cases, leaner, task-specific models are more than enough—and far more sustainable.
Putting the “S” in AI
AI holds enormous promise for tackling social challenges, from boosting education access to improving public health systems. According to Nature, AI could help advance 134 of the 169 UN Sustainable Development Goals (SDGs).
But poor governance can undo progress. The same reports warns that AI could hinder 59 of those 169 SDGs if risks aren’t properly addressed.
Many social risks are rooted in governance failures: biased recruitment algorithms, opaque credit-scoring systems, flawed datasets. To avoid this, we need ethics and governance to be embedded from the start—in AI design, in finance training, and in the broader public dialogue. We need a “just transition” for AI, just as we do for climate: one that ensures people and communities aren’t left behind as technology evolves.
No time to wait
AI can help investors move faster, see deeper, and act smarter. But real progress depends on how we wield the tools.
Institutional investors must take the lead—by investing in high-quality ESG data and innovative tools in their investment decision making process, while adopting robust AI governance frameworks, and ensuring all AI applications align with responsible investment principles.
This is a defining moment. Those who move now, and move wisely, will be on the front foot to drive sustainable financial and environmental outcomes.
