WASHINGTON, D.C. — ESG experts and sustainability professionals have long been optimistic about the potential for artificial intelligence to help solve the challenge of gathering the myriad data points companies need for compliance. This view has persisted as climate disclosure regulations like the European Union’s Corporate Sustainability Reporting Directive, the Securities and Exchange Commission’s currently-stayed rule and others in domestic and international jurisdictions continue to push companies to begin turning to the tool.
Generative AI served as the centerpiece for several discussions that took place at the Amazon Web Services Summit on June 27, which convened thousands of tech professionals. Sessions taking place at the Walter E. Washington Convention Center focused on how the growing technology can be used to comply with growing climate regulations or to create tools and insights to amplify social impact, showcasing the evolving range of use cases for the tech for ESG.
Marsal Gavalda, chief technology officer of ESG reporting software Clarity AI, who did not attend the conference, compared generative AI to the advent of the internet or mobile phones in its potential to shape society in an interview with ESG Dive later. Clarity AI’s reporting solution uses artificial intelligence to scrape company’s public reports and filings as part of its process, he said.
“AI, and generative AI in particular, is truly a transformative event in society,” Gavalda said. “Everybody's a little bit both excited and scared about what that's going to mean. We are more in the positive camp, in the sense that we believe that AI will actually be quite helpful … and we are mindful of the usages.”
ESG reporting remains a popular use case
While most global corporations have experience in reporting — with 96% of the S&P 500 and 81% of the Russell 1000 releasing corporate sustainability reports as of 2021 — the rise in regulations has placed an increased importance on the accessibility and reliability of that data. However, a KPMG survey of 550 business leaders at global public and private organizations with at least $1 billion in annual revenue found nearly half of respondents – 47% — still used spreadsheets for ESG reporting.
The influx of attention to AI — fueled by the rise of Chat GPT — has prompted a surge of investments into generative AI. Multiple recent surveys have shown an appetite among companies, and to a lesser extent institutional investors, for investing in the technology to aid its reporting plans.
Such data challenges led professionals like Joseph Mathews, a Fannie Mae technology architect, to one of the conference’s “lightning talks” Thursday, situated between a bevy of vendors showing off tech solutions aimed at the public and private sector.
With headphones on, Mathews and around 20 attendees gathered at scattered tables in the middle of the expo center, for a demonstration on how to use Amazon’s Bedrock generative AI product to streamline reporting and generate insights. The AWS representative showed how the tech giant’s product could utilize primary and secondary company supplier documents like utility bills and corporate spend and product composition to streamline reporting of scopes 1-3 emissions, increase user confidence in that data and further query the tool about trends.
After the demo, Mathews told ESG Dive that the secondary mortgage company is also among those currently doing voluntary sustainability reporting manually. However, he has been working with the finance team to identify a reporting framework and automated reporting solutions that are already on the market the company hopes to leverage, which brought him to the demo.
“There is a renewed focus on how this whole ESG reporting will work,” Mathews said. “Once the regulations are enforced … it's going to be a mad rush. So we are getting started to prepare for that already.”
Generative AI’s use for social impact increasing
The conference also featured sessions with Hurone.AI, a startup focused on bridging a healthcare access gap, and sustainable labeling organization Fair Trade. Taken together, the sessions offered a look at how artificial intelligence is transforming what is possible for impact-focused organization.
The ratio of oncologists to cancer patients in Sub Saharan Africa is 1-to-1000 and 1-to-2000 in Latin America, compared to 1-to-300 in the U.S., Hurone.AI founder and CEO Kingsley Ndoh said during a talk Thursday. Ndoh said a gap in clinical trial data exists for underrepresented populations, particularly from these regions, which could cause them to not get screened in time for certain cancers.
Hurone.AI’s goal, then, is to meet patients where they are and l use artificial intelligence to bridge that gap, by streamlining communication between oncologists and patients using a two-way messaging platform to connect oncologists with their patients for treatments and appointments. Ndoh said the startup’s pilot in Rwanda had a 92% engagement rate and led to a 60% reduction in off-duty calls for oncologists.
For now, the company is utilizing synthetic data to make up for some of those information gaps, but understands the importance of maintaining patients’ trust, Ndoh said. He said the company is looking to do that by working with the main stakeholders of that data.
“There’s a paucity of data in these regions … so, we're using what we have,” he said.
While Ndoh’s solution is centered around AI, Fair Trade USA — a nonprofit that certifies products as aligning with the fair trade mission — found ways to integrate AI into its production, proof of concept projects and for discovery, according to Olena Gomozova, a senior manager of engineering and solutions architects at Fair Trade.
Gomozova said the organization has found ways to use the tool both internally and externally, including primarily to translate user-generated content. Internally, the company is also utilizing generative AI to dig deeper into certain data and find compelling stories it wants to tell about its producers. Through their own implementation processes, the organization learned how to budget time for upskilling, allow for low-pressure testing and proof of concepts, stay up to date on updates in the sector and keep a human in the loop, according to Gomozova. In addition, the implementation process emphasized the importance of establishing organization-wide data literacy, she said.
“Data is everyone's job, and you have to have the full support from the organization and from organization leadership when you take it on projects like this,” Gomozova told the audience. “And then we need to ensure that we have the right people with the right skill sets.”
While the Amazon-hosted conference featured solutions that were predictably built on Amazon solutions. However, whether it is built on a Bedrock, Open.AI or any other architecture, Clarity’s CTO Gavalda said it is important that companies looking to break into the AI-powered ESG space should maintain unique data sets of their own and be mindful of their business strategy’s potential collision course with larger operators.
“Try to find something that is unique to your company,” Gavalda said. “Do not over rely on something like open AI, because if, basically, all you do is be kind of a [organizing interface for those platforms], then as soon as they launch some sustainability or some ESG sort of package, you would be out of business.”
Clarification: This story was updated to reflect the proper stylization of Clarity AI.
