Let’s not build a smarter hamster wheel
ESG, AI, and the inertia of perfect data
Today, there is no better applicability of the phrase, “Perfect is the enemy of good,” than in the space of carbon accounting.
I’ve made it no secret over the years that, in my opinion, the global disclosure schemes have unintentionally fostered an industry that prioritizes chasing ever more accurate carbon emissions numbers over actually doing the work.
Progress should have worked like this:
Measure.
Manage.
But instead, it seems to be:
Measure.
Measure better.
Keep measuring.
We’ve put ourselves on the hamster wheel of disclosures and started running, believing that one day we’ll catch the perfect data to…well, nobody knows what.
Now, as AI enters the ESG chat, the risk is that we double down on ever chasing those undefined outcomes.
Instead of using AI to help us act, we’re building ever-more-sophisticated tools to... measure better.
Don’t let AI perpetuate inertia…again
As companies mature in their disclosure workstream, data gaps are becoming apparent across hard-to-reach suppliers, uncooperative landlords, and non-integrated systems. AI can certainly help fill those gaps:
Collecting hard-to-find data: Generative AI agents can scrape unstructured activity data.
Estimating unavailable data: Machine learning models may generate proxy data for activities.
Enhancing auditing and reporting: Automation can be used for audit readiness or disclosure mapping.
These are clever uses, but they remain within the narrow swim lane of reporting and perpetuate inertia in the form of “if only we had just a little more data…”
Here’s the reality, which I’ve written about time and again:
For most businesses, carbon emissions are not universally material.
Chasing 95 %+ accuracy in Scope 3 data with AI isn’t helpful if your biggest material challenge is product safety, water usage, or workforce culture.
So yes, AI can help us report more accurately, but if it isn’t also helping us act, we’ve just created a shinier hamster wheel.
AI as the pivot to action
We don’t need a satellite to tell us it’s cloudy, just like we don’t need AI to push our carbon disclosures to the fourth decimal place. Yes, there are legitimate places it can help, but accuracy has a diminishing return on value.
That level of granularity isn’t ESG leadership. It’s a MacGuffin, a plot device that moves the story forward (and perhaps a significant part of our careers) but has little intrinsic value.
A better plot to follow is this:
AI can help us take better, faster, and smarter action on what’s actually material.
Materiality, as always, is where value lies.
A recent opinion piece in Fortune by Mark Cutifani, former CEO of Anglo American, and Johannes Teyssen, the former CEO of E.ON, called out the shifting nature of business and pullbacks from dealing with climate change.
Among their many strong points, this one stands out:
The situation and responses will be different for each sector and company. Companies should find where geopolitical, economic, societal, and sustainability priorities align, then choose actions that address sustainability and other issues simultaneously and contribute to resilience and competitiveness.
This is ESG. It’s always been what ESG is all about, and it's becoming what AI and ESG, together, can achieve.
So instead of just training models to assess and audit what we’ve already done, what if we used them to plan and optimize for what comes next?
Risk synthesis: Generative AI can analyze large volumes of industry and stakeholder reports, surfacing emerging risks and opportunities (even material issues) specific to your sector and company.
Climate foresight: Machine learning models can simulate the effects of climate change on assets, supply chains, and communities, enabling more innovative resilience planning.
Sustainable R&D: AI, paired with high-performance computing, can uncover novel materials or process improvements that reduce environmental impact during the design phase and beyond.
Operational efficiency: Automation and real-time analytics can reduce energy and resource use in operations, logistics, and asset maintenance.
These go well beyond reporting enhancements as strategic business accelerators that will get company leaders’ attention and drive real action.
Tools don’t lead. People do.
AI is a tool, just like carbon accounting is a tool, and just like sustainability reports are tools.
The timing of AI and the data modernization it will bring is fortuitous for sustainability offices if they recognize the opportunity to act.
Tools don’t drive change on their own. People do, and we need people with a new mindset to see what is required.
If we approach AI with the same mindset that turned ESG into a reporting exercise, we’ll get more of the same: smarter metrics, flashier reports, and very little progress.
However, if we shift our mindset to materiality, action, and long-term value, then AI becomes a force multiplier for better decisions, which would lead to more meaningful ESG disclosures.
That’s the call of this transformative force.
Let’s not build a smarter hamster wheel. Let’s get off it and use these tools to finally move forward.