Accelerate Sustainability with AI: A Practitioner's Perspective with Malte Øster

In today's data-driven world, AI isn't just on the map - it's the fast lane. From mining critical insights and tracking regulatory shifts to forecasting risk, monitoring employee health and safety metrics, and streamlining reporting, AI is steering the future of EHS, sustainability, and compliance. But as professionals take the wheel, what roadblocks and green lights lie ahead on the journey? This series of articles will help you navigate the future of AI in EHS.

Artificial intelligence (AI) has promised to revolutionize the way we process information by collecting, analyzing, and operationalizing vast quantities of data. It could, if wielded responsibly and ethically, provide solutions to some of the world's most intractable problems.

For sustainability experts, that is an appealing promise. Sustainability relies on data, forecasting, modeling, and analysis, and as these elements become more complex, the requirement for better technology tools to manage the burden becomes greater. AI, with its promise to augment human expertise with advanced pattern recognition abilities in a fraction of the time it would normally take, could indeed supercharge ESG and sustainability practices like reporting and regulatory analysis.

Despite what AI can theoretically accomplish, the reality is that we live in disruptive times, and AI isn't making predicting the future any easier than it has been in the past. To look at where we are today and what the future will look like, 3E sat down with Malte Øster, a senior advisor at SustainX in Copenhagen and an expert in EU legislation, sustainability strategies, and ESG reporting.

Impact of AI on Sustainability So Far

Øster highlighted the contradiction between the analytical power AI has given us and the uncertainty of the times in which we live.

“It's difficult to grasp both how fast or how slow things will grow and whether or not we've underestimated or overestimated the impact that AI will have,” said Øster. “I heard someone on a podcast say he thinks we're probably overestimating the short-term impact of AI but underestimating the long-term impact, and I think that puts it really well.”

For the last few years, AI futurists have been touting the revolutionary benefits of AI for sustainability, including optimizing energy use, reducing waste, and fine-tuning operations. So far, however, many of those promises have not yet manifested in the everyday work of the sustainability practitioner.

“AI has become more of a talking point over the last few years, but it hasn't yet changed a lot,” said Øster. “Of course, I'm using AI every day for my daily tasks to make me more efficient, but I don't yet see the fundamental changes to our work with reporting or with modeling and scenario analysis, which is where I think the use cases for AI are big.”

For Øster, human expertise remains a critical component of any sustainability solution that incorporates the powerful analytical capabilities of AI.

“If clients ask me about something in which I don't specialize, then I have more prerequisites to understand the legislation than the client does, so I'll research it,” said Øster. “One of my research tools is going to be AI, but it has a very literal way of interpreting legislation, and there is still so much that depends on human interpretation.”

The complementary relationship of human expertise and AI's data-processing abilities is likely to deepen over time. Øster provided an example of a request along the supply chain to ensure all providers are complying with human rights legislation. In a human-driven environment, that request might simply be passed from one person to another to fill in information, while an AI agent would likely approach it by opening an investigation that included mapping out relationships between suppliers. At the same time, the AI will only have access to information that is documented and available, while human agents might have tacit, expert information that is not documented in systems available to AI.

Beware of Garbage In, Garbage Out

AI is a powerful tool for extracting information, but to do that, it needs high-quality data. Indeed, poor data quality and data swamps can undermine even the most ambitious AI project. According to Øster, this is also a problem in sustainability.

“A lot of AI software companies are promising solutions in which their automation is dependent on their customers having perfect data already,” said Øster. “They promise one-click accounting in which you can upload all your data and then the model will sort through it, figure out the relevant invoices, extract the right data, apply the right factor, and you'll have a complete piece of climate accounting.”

Øster says this assumption is where the promise of AI can begin to break down.

“That would require the company to have all their data structured in the right way, but as a person working with climate accounting, that's probably 80% of what I do. Errors can creep in through extra zeros or missed commas, and that's where it makes sense to have some human common sense involved.”

Øster also sees opportunities to automate some of these processes even further and potentially reduce the amount of required human intervention, such as matching activity data from the company with emission factors, which could bring even more efficiency to the accounting process.

Short-Term Productivity but Long-Term Revolution?

For sustainability professionals like Øster, AI holds the obvious benefits of making everyday tasks easier and more efficient, such as generating text, doing quick research, or even translating emails from various suppliers around the world, therefore opening up more hours in the day to focus on high-value work. In the short term, AI is a powerful assistant to increase productivity and efficiency.

In the long run, however, AI has much more significant potential to create a revolution in the sustainability field.

“The role of AI in scenarios and forecasting is where the real business case becomes quite salient,” said Øster. “Insurance companies and banks need to be able to predict risks for things like insurance pricing. They're starting to use AI modeling of their risk appraisals to make their predictions more precise, and they're discovering they've underestimated the environmental risks they were exposed to. That kind of ability will fundamentally change the business case for sustainability, and it will make companies take sustainability seriously, if they haven't been doing it already, because it's good business.”

Achieving this brighter future for sustainability is not, however, a straight path, and obstacles remain. Regulatory action and financial incentives will be critical components in the task of ensuring climate risks become a central consideration for strategic planning in the world's financial institutions.

As is the case with many ambitious technology projects, AI in sustainability will also need strong digital foundations to make a real difference.

“We'll need more data to be able to model, and we'll need more computing power to forecast some of these things,” said Øster. “As we know with things like weather forecasting, you need quite big data sets to make precise predictions, and you will eventually need quantum computing power to process that much data.”

Cost of Shift to AI

As is well known by now, the progress of AI is coming at an environmental cost. The more data you have, the more computing power you need, and the more energy you consume. From water consumption to greenhouse gas emissions created by massive data centers, AI is the root cause of many of the sustainability problems it's been tasked with solving.

On the brighter side of the paradox, Øster sees the possibility of greater awareness of sustainable solutions.

“The obvious answer is more renewables,” said Øster. “One of the good things about the world waking up to the fact that we need an infinite amount of energy is that we need to get it from an infinite energy source, such as the sun or the wind, instead of taking even more finite resources like oil and gas from the ground. That's a key consideration for a lot of the new data center projects around the world.”

Like any technology revolution, AI is going to change the way we work. In the short term, that means obvious benefits in efficiency and productivity.

“I would hope that AI can remove all of the tedious work around stuff that I'm doing related to documentation, process reports, and so on,” said Øster. “Then I could really focus on the things that make an impact, like talking with stakeholders, and not just typing things into my computer.”

The reality, of course, is that AI is likely to replace not only tasks but also a number of the people who perform those tasks, especially junior people.

“I don't think my position necessarily exists as it does today in 10 years,” said Øster. “Making more information freely available online and creating an AI specifically tailored to a company's preferences could replace a lot of people.”

Human specialists, however, do much more than simply provide expert information. As Øster points out, they also provide authority and responsibility for the work they perform.

“I also take on some legal responsibility when I tell clients if what they're doing is good and if it lives up to the legislation. Clients can't simply say, 'an AI told me to do this,' so I take on some of the risk and the legal responsibility for those things.”

On a final note, Øster emphasized that the power of AI to generate information could potentially be both a curse and a blessing.

“It's going to be important to be able to filter information and to know what to look for,” he said. “With the rise of text generation, there are a lot of buzzwords and AI-generated text that doesn't provide much value. I think if you're able to cut through that noise to focus on the things that are truly material and the information that's relevant, you're going to go a long way.”

AI is a powerful tool that will fundamentally change the information paradigm in which we live, and it holds considerable promise to give sustainability the power it needs to make a real difference in the fight to protect the environment, reduce global warming, and distribute equitable social justice around the world. As for humans, though our roles and responsibilities might change, our expertise, insights, and priorities will continue to be critical components of the new paradigm as we apply the power of AI to protect the planet and future generations.