September 16, 2026, (Inside AI) — Robert Eccles, a visiting professor of management at the University of Oxford's Saïd Business School, has a message for corporate America: stop playing defense on sustainability and start using artificial intelligence to make environmental risks financially legible.
In a new book co-authored with Columbia Business School accounting professor Shivaram Rajgopal, Eccles argues that AI can bridge the gap between sustainability reports and the hard numbers in corporate financial statements. The book, titled Making Sustainability Financially Relevant: A Man/Machine Collaboration, shows how large language models can map environmental, social and governance data directly to income statements, balance sheets and cash flow statements.
The timing matters. Companies face mounting pressure from Republican state attorneys general and Washington policymakers who have attacked ESG investing as a political agenda. Yet many firms continue issuing sustainability reports. The missing piece, according to Eccles, is a clear financial translation.
"For a long time, nobody really knew what ESG was. And then they kind of thought it was philanthropy and then sort of didn't like it," Eccles said in an interview. "And then it's about the material issues, you know, sector specific that matter to value creation."
The book's central claim rests on a practical experiment. Rajgopal analyzed ExxonMobil's 10-K filing using ChatGPT and a set of prompts. The process took roughly 100 hours. Eccles then built what he calls a black box model, essentially a massive prompt, to streamline the analysis. The result: the time an expert needs to estimate the financial consequences of material risks dropped from 100 hours to 30 hours. Since the book was finished in the spring, Rajgopal said AI advances have probably cut that number to around three hours.
The approach uses the Sustainability Accounting Standards Board framework as a starting point. It compares a company's own disclosed risk factors and footnotes against SASB's sector-specific material issues. The AI then maps those issues to specific line items in financial documents.
"With ExxonMobil, we basically took the data in their 10K and we can show that pretty much all of the SASB issues are addressed," Eccles said. An Exxon representative declined to comment.
The organizational implications are significant. Eccles writes that most companies cannot answer a basic question: who owns the relationship between sustainability risk and financial exposure? The answer, in most cases, is no one.
"The obvious solution is to have the CSO report to the CFO," Eccles said. "The CSO has to be able to understand financial stuff and the CFO needs to understand sustainability stuff."
That structural gap has left companies vulnerable to political attack. Eccles traces the ESG backlash to both ends of the political spectrum. The left, he said, initially pushed companies to make the world a better place without distinguishing material financial risks. The right then weaponized that confusion.
"Companies kind of created the opening to be attacked," Eccles said. "They confounded value creation with making the world a better place."
He points to a letter sent by Republican state attorneys general to the Big Four accounting firms, including Deloitte. The letter, led in part by Nebraska Attorney General Mike Hilgers, accused the firms of supporting external climate reporting standards to profit from them. Eccles noted that Nebraska itself received $307 million from the Environmental Protection Agency under the Biden administration, the highest per capita award in the country. Much of that funding went to agricultural data and grants programs, with Deloitte assisting on climate and financial controls.
"They express these concerns that they have no data whatsoever that any of these problems that they're concerned about exist," Eccles said. "It's kind of like getting this in the public eye, get some attention. It's theater."
A representative for Hilgers declined to comment. Deloitte representatives did not respond to requests for comment.
The AI angle extends beyond efficiency. Eccles and Rajgopal argue that human expertise remains essential for verification. The book's subtitle, A Man/Machine Collaboration, reflects that division of labor. AI accelerates the mapping and estimation. Humans check the outputs and take responsibility.
"You've got all of these arguments in the academic literature that users should disclose AI, and how you can only use it together," Eccles said. "It's like, no, everybody lies about it anyway. And so what do you care if the human being puts their name on it?"
That view aligns with a broader shift in AI-assisted professional work. Tools like ChatGPT and other large language models are increasingly used in accounting, law and journalism for document review and pattern extraction. The value lies not in replacing experts but in compressing the time required for tedious analysis.
Read: Anthropic Launches Claude for Financial Services, Challenging OpenAI in Finance AI
Eccles remains optimistic despite the political headwinds. He sees the current moment as a wake-up call. Companies, he argues, should stop playing defense and start explaining how sustainability issues affect financial performance.
"I think companies need to quit playing defense. They should go on offense," Eccles said. "It's not like they need to be screaming at Trump or Republicans, just like, here's our narrative. These are the sustainability issues we think matter to value creation."
The book arrives as AI adoption in corporate finance accelerates. According to industry surveys, a growing share of CFOs are piloting AI tools for risk assessment and reporting. The challenge remains governance: ensuring that AI-generated financial estimates meet audit standards and regulatory scrutiny.
For now, Eccles and Rajgopal offer a template. Use AI to surface material issues. Map them to financial statements. Put a human name on the final call. The approach may not end the ESG wars, but it could give companies a defensible position.
