The recent explosive growth in the development and deployment of artificial intelligence (AI) has dominated conversations in 2023. Part of those discussions have centred on the transformative potential (or indeed, reality) of AI on the climate tech industry. With this in mind, we ran a number of Climate Tech roundtables over the course of 2023 with our good friends at Undaunted. The last of these focussed on the opportunities and challenges in using AI to assist in the net-zero transition. Attendees ranged from large corporates and industry experts to start-ups and other sustainability advisors, and discussions covered issues such as:
- The difficulty in defining AI – As with discussions about all implementations of AI, it is important to look beyond marketing hype and determine whether a solution does include AI, particularly as this may determine whether it is caught by particular AI legislation. As the UK government’s White Paper on Regulating AI acknowledges, “there is no general definition of AI that enjoys widespread consensus” (please see here for our publication on the White Paper). The White Paper defines AI by reference to two key characteristics – 'adaptability' and 'autonomy'. By way of comparison, the EU, which faced much criticism for the wide definition in its original proposal for its AI Act, recently agreed to align its definition with the OECD definition of AI Systems. Noting the importance of a practical AI definition, and the criticality of assessing its application in order to ascertain the regulatory position, good practice dictates that many AI related obligations (around transparency, fairness, avoiding bias etc.) are equally relevant whether or not AI is used.
- The importance of understanding the datasets – there was common acknowledgement of the importance of understanding a company's data sets, and the organisation of those data sets, as a pre cursor to the effective deployment of AI solutions. Whilst powerful if implemented in the right environment, there was wide spread recognition that AI relies on robust and effective data governance.
- The importance of building trust – for AI adoption to increase, trust is needed in the AI products. From questions around greenwashing risk, to their environmental impact and the accuracy of outputs, potential customers need information to allay these concerns. Transparency and explainability are therefore key. This is particularly the case in the Climate Tech space, where the upsurge of attention in recent years may create increased customer nervousness of inflated hype.
- Keeping on top of a changing regulatory landscape - As 2024 reaches its conclusion, we’ve seen big headlines in both the AI and Climate Tech worlds. COP28 drew to a close on 13th December (see our COP28: Impacts for business - Slaughter and May Insights for more information), while the EU finally reached political agreement this December on both the Sustainability Directive and AI Act. Regarding the latter, this is the “first-ever comprehensive legal framework on Artificial Intelligence worldwide” which considers environmental issues as part of its many provisions. For example, high-impact general purpose AI models with systemic risks will face a number of obligations, including having to report on their energy efficiency (see our blog). Given the pace of regulatory change, it is important that organisations keep abreast of their evolving obligations.
- Embracing the wide range of opportunities AI presents – while there has been much written on the opportunities AI presents (see for example our ESG section from our Regulating AI Series), much of this has focussed on its potential to manage large, complex, data sets for ESG reporting purposes. However, AI use cases are diverse, resulting in a variety of ways in which AI can help in the transition to net zero. Opportunities discussed ranged from data centre optimisation to smart plugs, carbon off-setting backed by reliable data and even the use of AI to assist overstretched ESG teams increase their productivity.