Sustainable Development Goals (SDGs) – an ambition too far?
by Petra Parizkova, Emily Williams
Due to increasing stakeholder expectations, more and more companies are reducing their emissions of greenhouse gases. Net Zero is quickly becoming the standard as Europe strives to be the first climate-neutral continent by 2050. Corresponding sustainability strategies require companies to collect, process and leverage growing amounts of data, especially in the supply chain; aspects which AI applications seem predestined for. So, what will we still need people for in the future?
This article was originally published in German by CSR-News.
By Daniel Silberhorn, Senior Advisor for ESG & Sustainability Transformation at SLR
‘The human factor’ – I was recently invited to a phone call on this topic by an internal working group. What if, in the future, gathering, calculation and reduction of greenhouse gases are increasingly done by AI? This is the question on the minds of the teams who help companies to calculate and reduce their carbon footprint.
A question that has been on the minds of many experts from a range of industries since ChatGPT was published in late 2022. As early as February 2023, that AI chatbot reached 100 million active users per month, taking the throne of the fastest-growing application of all time – and it is only one of a growing range of so-called generative AI applications that can produce various types of content, including text, imagery, audio, and synthetic data. Observers agree, consequences will be huge.
In fact, many professions will probably experience accelerated change through AI. ChatGPT developer OpenAI itself estimates that around 80% of employees in the USA are in jobs in which at least one task could be completed more quickly by AI. Accountants, mathematicians, programmers, interpreters, writers, and journalists are all affected. What does this look like in practice for one of the most important current corporate sustainability topics - the targeted reduction of greenhouse gas emissions, and in particular, their Net Zero strategies?
The term 'Net Zero' first appeared in a 2018 report by the Intergovernmental Panel on Climate Change (IPCC). According to this report, all countries need to reduce their carbon emissions to Net Zero by 2050; this is the only way to limit global warming to 1.5°C compared to pre-industrial levels.
Companies can achieve Net Zero by reducing their own greenhouse gas emissions as much as possible and offsetting the remaining emissions by investing in projects such as sustainable forestation or renewable energy. In the meantime, Net Zero is increasingly becoming the standard for corporate climate strategies. Many of the steps required for a carbon footprint calculation and a Net Zero strategy are laborious human efforts supported by digital tools. It ranges from data collection and data cleaning to the calculation of overall emissions, based on which targets and concrete roadmaps are developed for implementation. Then, the final step is to motivate people to actually change their behaviour – for example through more climate-friendly mobility.
It seems tempting to take advantage of the capabilities of current AI applications. However, AI applications devour large amounts of energy and are therefore themselves a climate-impacting factor. The development of ChatGPT-3 is said to have caused an incredible 552 tons of CO2. According to the MIT Technology Review, a single AI model can cause as much CO2 as five cars over their full lifetime.
With that in mind, if you are to assume (and hope) the AI uses green electricity, there are promising areas of application. Its strength lies in its ability to collect large amounts of data, process it quickly, examine it for patterns, and generate options. On the one hand, AI is therefore predestined for the analysis and modelling of emission data. Appropriately trained programs can theoretically quantify the CO2 footprint of companies and identify opportunities for savings – especially where emissions are high. AI-supported models can also be useful for forecasts and scenario analysis by simulating different scenarios and predicting the effects of measures on CO2 emissions.
Equally, AI is seen as potentially valuable in the future to help optimise the use of resources, with a view to reducing greenhouse gases: for example, in energy consumption, logistics or production processes. This extends to the employees’ behaviour, for example to reduce personal emissions via sustainability apps. AI can also play to its strengths in the monitoring of emissions, and it will probably also support transparent and increasingly accurate sustainability reports in the future.
So, can humans all retire soon in these areas? Leave everything to artificial intelligence? I would like to say: Of course not. There are several clear reasons against it. However, at least one of them could become superfluous as the quality of artificial intelligence improves further.
During its first presentation in February 2023, Google's AI chatbot called Bard made a spectacular mistake in front of the world, and simply made up information. The tricky thing is that AI results are often very convincing. However, the results are only ever as good as the data that an AI accesses - and how it handles it. This can generally be solved through training and better data.
The same applies to the complexity and uncertainties associated with the implementation of Net Zero strategies. The level of complexity is often high and varies from industry to industry and from company to company. Data must be collected from all business divisions and stages of the value chain, checked for quality, and standardised for use. Where information is inaccurate or even missing, sustainability managers must work with estimates that stand up to scrutiny.
It becomes particularly complex considering the increasing responsibility for entire value chains, which can include a large number of industries and regions worldwide. Close operational cooperation with suppliers and partners is necessary already when collecting emissions data. This requires careful relationship-building to enable trust and reliable collaboration. In addition, Net Zero strategies require long-term planning of measures and investments. Assumptions and estimates for various factors are incorporated here, and these include technical progress, regulatory developments, and the expected price volatility of CO2 emission allowances.
In practice, our carbon experts find that the availability and quality of data often represents a major hurdle. This can and could lead to incorrect results that would be difficult to interpret when using AI in the future. What exactly goes into the calculations, how does the AI fill any gaps?
In addition to the topic of transparency, these questions also refer to the question of practical and ultimate responsibility. A key question of AI, which has not yet been conclusively answered in the case of autonomous driving or, to an even greater extent, in military use. What level of control by human experts is guaranteed - and who is responsible for wrong decisions?
For the foreseeable future, it will therefore remain important to know the limits of AI-supported models and forecasts and to use applications correctly. They can only succeed with human expertise, and just as the mention of employees and suppliers above already refers to the human factor, there is only success if those affected are involved in planning and implementation.
For example, a global manufacturer of showers approached our company to develop a robust internal and external approach to its sustainability. Being protected against greenwashing allegations was crucial - a topic that is currently on many minds. After all, authorities and legislators are actively cracking down on greenwashing, see for example the new EU Green Claims Directive.
The implementation of sustainability is a management task that requires changes to corporate structures and processes – and often even to corporate culture. Companies must therefore promote acceptance and commitment among employees to ensure successful implementation.
In the first step, we therefore developed a shared understanding of sustainability among key executives, including the specific level of ambition. A storytelling workshop and competitive analysis helped produce in an authentic narrative. This provides the communicative basis for leadership and change management: What are we doing? How do we do it? And why are we doing all this?
We advised to form a strong cross-functional coalition that included all the departments that contribute to and will implement the strategy. During the development of the strategy, we spoke to numerous internal stakeholders. However, it turned out that one department was initially left out by accident. This department later turned out to be sceptical and only reluctantly cooperative when the sustainability strategy was to be rolled out. The solution: discussions and another workshop in which we looked at the perspective of those involved and addressed their professional needs. A key point: strategy is always co-creation, with those responsible for sustainability playing a central role.
Part of the implementation agreed on was an internal communication campaign, which included an internal Sustainability Day for which we provided advise. Internal stakeholders from production to the sales team were informed and provided with suitable materials for their work. An internal network of sustainability champions can also anchor the topic in the culture; at the same time, other cultural aspects come into view, such as onboarding, bonuses, governance, and internal media.
In fact, many transformation projects fail due to a lack of communication, losing sight of the human beings involved – and sustainability often represents a major transformation indeed. John Kotter’s Change Model attributes the key mistakes in change management to failed communication and stakeholder engagement: not enough awareness of urgency is created, insufficiently strong internal coalition, no clear vision that sets the direction, and too little communication about the vision.
Net Zero strategies also require cooperation with external stakeholders such as customers, suppliers, investors, and authorities, as well as associations and alliances such as the UN campaign 'Race to Zero' or 'Transform to Net Zero'. The supply chain has recently received particular attention: In the past, communication with suppliers was mainly via contracts negotiations - what is supplied under which terms and conditions. In times of the Supply Chain Act in Germany and the discussion about the Corporate Sustainability Due Diligence Directive (CSDDD) at European level, however, this now requires an open dialogue and collaboration to implement measures to reduce CO2 in the supply chain. As part of the Scope 3 emissions, these are increasingly coming into focus.
These considerations make it clear: the development and implementation of sustainability and Net Zero strategies go far beyond working with data. It's not possible to just dump figures into an AI black box and get a perfect strategy that works and is accepted. It’s about people. The human factor.
On the way to strategy, AI is a potentially powerful tool that can bring great benefits if used according to transparent and clear standards. However, the human factor is and remains central when companies successfully embark on their path to net zero emissions.
by Petra Parizkova, Emily Williams
by Josh Kahn, James Balik-Meacher, Luke Bertini
by Richard Jardine