Like all knowledge workers, sustainability professionals are trying to navigate the accelerating deployment of artificial intelligence along with its ravenous appetite for energy, water and other critical inputs.
Another important question is less considered: What can we not afford to turn over to AI? Where should a human-in-the-loop become a human calling the shots? Here are three instances:
Don’t let AI dictate your level of sustainability ambition
AI optimizes against the value model it inherits. For most companies, that model is the financials as currently structured — where carbon intensity doesn’t show up in sourcing decisions, water risk isn’t booked as a cost and supply chain fragility is invisible until it becomes a crisis.
AI agents trained on that model won’t just overlook sustainability value; they will optimize against the identified business objective to the exclusion of other variables and time horizons. Ask AI to reduce procurement costs, and it will provide a solution for today, without considering supplier performance, impact on Scope 3 emissions or how that supplier shows up in its community. AI is now being used by teams across real estate, capital allocation, manufacturing and product development in exactly this way.
Telling AI to “also consider sustainability” is not the solution because instructions get retrained out in the face of optimization. We’ve been talking about the fix for 30 years: embed the value of what’s currently externalized directly into the business model.
Someone has to decide what goes into the model before it runs. Embedding sustainability considerations into the company’s business model — translating risk and opportunity into revenue — is the fundamental task of sustainability professionals.
Don’t let AI substitute for forward-looking strategic business insight
AI is a pattern-recognition engine. It identifies what has happened in the past and projects it forward. It cannot read across systems with no shared language, no common data structure and no historical record of intersection — which is exactly the work sustainability professionals do every day. We connect ecological and social systems that operate on different timescales and metrics than the financial system, and identify where they’re going to collide with the business model before the impact shows up on the bottom line.
Carbon emissions now show up as stranded assets on oil and gas balance sheets and a tariff cost at the EU border. Water scarcity has rerouted semiconductor supply chains and delayed data center deployment. Forced labor exposure has triggered import detentions and bans that stalled inventory for apparel and auto companies at U.S. ports.
None of these were surprises. They were external signals on a trajectory to intersect business strategy, and they were visible long before they hit the income statement, at least to anyone trained to read them.
Don’t deploy AI without a sustainability professional in the room
Externalized costs tend to become corporate crises. Carbon emissions that companies failed to book as a cost two decades ago are now showing up as stranded assets and tariff exposure. Water historically treated as free is now a supply chain limitation. AI is already generating the next round of crises.
Klarna replaced roughly 700 customer service roles with AI between 2022 and 2024 and publicly celebrated the labor cost savings. By 2025, the CEO was rehiring — not because the AI failed at its task, but because the optimization function didn’t include the very real value of human customer experience.
When xAI selected Memphis for its Colossus supercomputer, the 100,000-GPU cluster came online in 122 days inside a retrofitted factory. The public only learned of the project the day the Greater Memphis Chamber of Commerce announced it in June 2024. To make up for power-grid shortfalls, the company installed dozens of unpermitted methane gas turbines on-site — in a majority-Black community in South Memphis.
The facility is now facing Clean Air Act lawsuits and injunctions. This is a textbook risk-management failure that a sustainability professional in the deployment conversation would have flagged before the first turbine was installed.
These aren’t abstract moral concerns. They are business design decisions with material consequences — for resilience, for regulatory exposure, for brand, for the supply chains and communities the business depends on.
The sustainability professional’s job has always always been to embed sustainability into the company’s operating system — the model the business uses to allocate capital, evaluate risk and decide what’s worth building. We got sidelined measuring impact and setting aspirational targets, instead of shaping the decisions that create those impacts. AI is now executing against those unchanged business operating models at machine speed.
It’s time for sustainability professionals to double down on the work we are trained for: surfacing what isn’t yet priced before the market forces the correction.
At the NYU Stern Center for Sustainable Business, we’re building that capability into the next generation of business leaders. Sustainability isn’t a standalone subject; it’s competitive intelligence: the discipline of reading what the financial system can’t yet see and translating it into strategic action before the market forces the issue.
Students who come through our program leave knowing to ask, What is this company dependent on? How is that changing? and What does it mean for financial performance? That’s the systems view that businesses need, and only humans can provide it.