Environmental, Social and Governance (ESG) has, for some time now, been a huge subject for accountants. It features in ICAEW’s 2020-2030 strategy as ‘Helping to achieve the Sustainable Development Goals (SDGs)’. 

Increasingly, ESG and technology are converging. The Venn diagram of ESG and tech is showing more and more overlap, both in terms of the opportunities that are presented in the space and also in the risks that technology poses in meeting key ESG objectives. As COP 28 approaches, what is the role of technology in the world of sustainability?

Good reporting requires good data

Natalie Pullin, Sustainability Lead at Fujitsu, explored some of the major features of working towards net zero in her keynote at the ICAEW Annual Conference. Many of them come back to the importance of good quality, granular data. 

Pullin observed that as organisations’ desired approach to sustainability shifts from being more compliance led to more business-decision led, this will require a substantial change in terms of the level of information that is captured and analysed. 

While more immediate ESG disclosure requirements are focused on very high-level figures, most notably around carbon emissions and gender pay gaps, in order to really drive change and improve the fundamentals of how businesses operate, ESG data needs to be captured at a transactional level so as to pinpoint the areas that are both most in need of change, and also potentially will deliver the best return on investment. With this, the session from Auditel outlined its five key principles of carbon accounting:

• Relevance
• Completeness
• Consistency
• Transparency
• Accuracy

An important observation made by Auditel was also the contextual piece – being able to articulate what the data means, and what ‘good’ looks like.

Technology opportunities in ESG

As noted in our previous article, automation cannot be an afterthought when looking to deliver reporting on ESG metrics. Given the volume and complexity of data involved, it will be essential to automate as much of the data collection and processing as possible, to avoid it becoming its own industry and sucking up valuable time among already stretched reporting and sustainability teams. MHA shared how its automation also includes controls to flag inaccurate or insufficient data as a way of ensuring that ESG reporting meets required standards and does not lead to ill-informed decisions.

Other potential areas of opportunity that were mentioned throughout the day included digital twins, which can allow, for example, the virtual development of new products or the simulation of wear and tear on existing physical assets. This allows organisations to make more effective business decisions without having to waste real-world resources.

Smart IoT sensors could monitor air quality, energy usage, heating and cooling in buildings, which is a key element of sustainability reporting. Using workflow and process mapping tools could identify where different elements of the supply chain touch different ESG objectives.

Platforms and SaaS solutions are moving away from spreadsheet-based solutions to dedicated cloud-based platforms, which should lead to more accurate and more efficient tracking of ESG targets.

The Unanswered AI Questions

As the world gets more excited about the use of tools such as ChatGPT, this is leading to further questions about the environmental and social impact of this new wave of AI. 

Large Language Models (LLMs) such as those used by ChatGPT consume significant amounts of energy to power the servers, and water to cool them – estimates suggest that ChatGPT requires as much electricity as a large town, and as much as a 500ml bottle of water for every conversation. The challenge is how we manage this effectively, as such an increase in energy and water consumption is not necessarily compatible with achieving the UN Sustainable Development Goals (SDGs). 

It was reassuring that attendees at the Annual Conference were increasingly keen to understand how organisations such as Microsoft and Fujitsu are considering this risk, but unfortunately there are currently no easy answers. 

MHA’s view at the conference was that the emissions generated through the use of AI to support ESG initiatives should not, ultimately, outweigh the benefits derived from doing so – an opinion supported by Pullin. “There is a meeting in the middle – a need for proper, well thought through IT strategy with sustainability embedded into every function of an IT organisation, while being mindful of energy used in AI models,” Pullin said. 

Critically, Pullin believes that more could be done around how we use technology more efficiently and the waste of legacy solutions: “We all play a part in building new stuff, storing new content, without always thinking about removing/replacing/improving what exists already.”

From a societal and governance perspective, AI also poses significant challenges, many of which are yet to be fully addressed. There is much research to suggest that AI, when used to support decision-making, can perpetuate biases in society. Many speakers at the conference asserted therefore that governance and ethics will play a crucial role in ensuring we utilise AI safely and responsibly; an organisation’s leadership will also need to show, not just tell, their employees and other key stakeholders how to embrace AI and technology in a sustainable way.