AI sleuthing out ESG disclosure deficiencies?
– Companies should keep an eye on how AI-powered platforms are evaluating their ESG disclosures.
– So far, it’s unknown how much of a “black box” these platforms are – in other words, it might be challenging to determine how accurate these platforms are in assessing corporate disclosures (like it is for so many other ranking & rating services).
So a while back, I excerpted this from a Financial Times article:
But it is remarkable just how much data already exists. More striking still, a global army of entrepreneurial geeks is racing to create platforms, many of them open source, to parse this information. This roster includes establishment names like Goldman Sachs, start ups like Arabesque and academics. Two examples emerged this week: a team at Cambridge university is building an open-source system to track ESG risks in sovereign bonds; their counterparts at King’s College London are scanning companies’ AI platforms for hidden ESG problems.
This makes data access more democratic at the same time that new digital platforms are also enabling once-powerless voices to unite in protest more effectively. The #MeToo movement shows how this can, sometimes, change power dynamics. And while grassroots protests can be capricious, the combination of these trends is changing the psyche of corporate boards. So while corporate leaders used to view ESG issues as an optional “nice to have”, they increasingly feel that it is dangerous to ignore them. ESG blind spots can create reputational damage and cause companies to lose employees, customers, investors or regulatory licenses. Talking about it is a tool of risk management for corporate boards.
And now we have this article from Bloomberg Green – here’s an excerpt:
Hoepner and his team—called GreenWatch—are using artificial intelligence to parse media statements, websites and other corporate communications for sustainability claims by 700 global companies. They then compare those pro-environmental statements with a company’s carbon footprint to see whether it is cutting carbon emissions by 7% year-on-year in order to meet 2050 net-zero targets.
In their analysis, GreenWatch ranks companies based on whether their sustainability claims hold up. The group also focuses on what companies say about climate change: from whether they merely agree that global warming is something that needs addressing to stating outright they are climate leaders. The latter, GreenWatch said, can boost the likelihood of corporate greenwashing
The group ultimately places companies in one of several categories: green leaders, hidden green champions, green incrementalists or potential or probable greenwashers.
GreenWatch’s early findings show that there’s a high likelihood of greenwashing in 95% of the statements they analyzed from communications companies, which includes telecommunications and media businesses. Meanwhile, more than 80% of statements from corporations in the industrials, materials and consumer discretionary sectors have a high likelihood of greenwashing. But less than half of the statements from energy companies had a likelihood of greenwashing, GreenWatch found.
Geographically, 84% of statements from Japanese corporations had a high likelihood of greenwashing, the most globally, GreenWatch said. The U.S. followed with almost 75% of statements.
GreenWatch isn’t the only group using AI to sniff out potential greenwashing. Researchers at the University of Zurich and Germany’s University of Erlangen have analyzed corporate disclosures of climate risk—such as those recommended by the Task Force on Climate-related Financial Disclosures (TCFD)—to see if companies are cherry-picking information they report.
Their model, dubbed ClimateBert, “comes to the sobering conclusion that the firms’ TCFD support is mostly cheap talk,” the researchers said. “From our analysis, we conclude that the only way out of this dilemma is to turn voluntary reporting into regulatory disclosures.”