top of page
  • charlottemac1

Bridging Data Gaps and Ethical Frontiers with AI

Numerous panels at COP28 spoke on the importance AI. Data gaps remain one of our biggest barriers to sustainability. You can’t manage what you can’t measure. For example, only 126 countries reported on their populations using safely managed drinking water services, therefore we lack information regarding the status of 67 countries.

In 2022, UN member states reported their status on approximately 8.5 out of the twelve targets for SDG 6 (Clean Water and Sanitation), and only 50% of its members reported on nine or more targets. In order to mitigate the risks of climate change, let alone succeed in our mission for sustainability, we need to close these data gaps. Without this knowledge, we are more blind to the real-time effects of the climate crisis which can lead to poor risk mitigation strategies, ineffective circular economy planning and beyond. In the face of this gargantuan task, many have turned towards artificial intelligence (AI) for solutions.


For all the potential AI has, we are still in a relatively primitive stage for its applications. We are primarily examining the AI tools which have far larger capabilities than generative AI platforms like ChatGPT, although even these may have a role in transitioning to a more sustainable world. Numerous panels at COP28 spoke on the importance AI has in progress towards, its cross-sector applications and most importantly on how it should be used to fill our data gaps in terms of the SDGs.

Countries Reporting on SDG 6 Targets

Percentage of UN Member States Reporting on targets for SDG 6. Data adapted from: UN Water (2023) – Data Gaps, SDG 6 Data

Examining the data that the UN released for countries reporting on SDG 6, the only target that has recorded 100% of applicable countries reporting, is target 6.a.1 (amount of water and sanitation-related official development assistance that is part of a government-coordinated spending plan). All of the 139 countries that are eligible to receive official development aid (according to the DAC list of ODA Recipients) reported on this.


Removing the targets which are applicable to fewer countries (6.5.2 and 6.a.1), the average amount of UN member states reporting on all SDG 6 targets is 72.66%.


We are missing more than a quarter of the full story for SDG 6, affecting mitigation and risk strategies for natural disasters as well as solutions for sustainability.

How Does AI Fit into This?

AI was a significant point of focus for many COP28 events across the zones, with the scope of its reported uses stretching from planning resilience and mitigation from the effects of climate change, to enhancing agri-food climate systems and other climate solutions. The speed that AI can process data makes it integral in moving forward at a faster pace for data-gathering and analysis to better inform our decision making and planning of positive climate action.


Satellite imagery powered AI data processing has huge applications for improving sustainable development goal strategies, reporting and efficiency. For the agriculture sector, satellite imagery has been linked to mapping areas and regions that are good for farming practices, monitoring the use of pesticides and fertilisers and also plays a part in mapping areas at risk from natural disasters such as forest fires.

AI-powered irrigation monitoring, such as that provided by Seabex, can be used to monitor water-soil content and provide insights to prevent overwatering and reduce water waste (Mestiri 2020). Seabex is an agri-tech start-up that uses AI to deliver agricultural solutions for improved crop yields and reducing water usage, amongst other intelligent solutions. Reducing water usage is critical as the agriculture sector is responsible for approximately 70% of natural water withdrawals (Khokhar 2017). Reducing water waste for this sector plays a key role in agri-water strategies.

In terms of risk mitigation and adaptation, Microsoft has partnered with Planet to deliver AI solutions that analyse images of Earth in real-time from satellites. The initial step in their strategy is to first locate all human structures across the world. Understanding where everyone lives is vital in mitigation planning; without this knowledge, our planning and predictions are filled with data gaps. The data will help inform and predict migration patterns in correlation to weather patterns, natural disasters, geological trends and sector impacts. For example, long-term severe heatwaves may result in mass migration for a percentage of a population across a country.


Knowing where people are going to or from in terms of permanent or semi-permanent residence will help inform decisions relating to the WASH sector, utilities, healthcare, food security and beyond so that infrastructure, support and capital can be effectively deployed.


Many countries in the Global South face a far larger data gap barrier to those in the North where often 10-15 years of geological and sociological data are missing. A large proportion of these countries are on the front line of the climate crisis and suffer the magnified effects of climate change. Generating and obtaining this data is vital in fairness for climate action, implementing climate solutions, planning risk mitigation, adaptation and in climate risk management, to name a few.


AI clearly already has a vital function across sectors in improving sustainable performances but could the applications of AI prove useful in the transition away from our fossil fuel dependence? There are reports of AI tools being used by big players in the fossil fuel industry, such as Shell, to accentuate sustainability initiatives. In fact, machine learning has been used by the fossil fuel industry since the 90s. Unfortunately, in some cases AI-fossil fuel partnerships have actually correlated to increased production of fossil fuels and has been linked to data analysis to prevent mis-drilling under the guise of these sustainability initiatives (McLennan 2023). Whilst reduced mis drilling is certainly not a bad thing, this reduces the transparency of AI use, creating a data bias in the use of AI in its relationship with our environment.


This brings to light a segment of ethics for AI. How do we regulate AI to ensure its powers and capabilities aren’t abused for negatively impactful activities?


Some ethical concerns surrounding AI relate to data and privacy on a personal level. However, we are examining the larger scale applications that are often hard to pin down. First, we must consider AI as a technology or a process, rather than a product, as AI provides the “how”. The products or projects that AI can produce as a consequence need to be regulated and reported on to ensure transparency and reduce misuse.


Regulating AI as a technology is tricky. Consider a laptop, which may not be particularly regulated as a technology. However, the negative actions someone might use a laptop for can be regulated. Regardless, regulating AI technology will require transparency, openness and good governance. Larger organisations, such as the Ministry of Defence, GCHQ, and the NHS have established strategies for AI and guidelines for assessing safety and ethical robustness.


In terms of regulating AI in sustainability, the EU AI Act, expected to come partly into force by 2026, is aimed at ensuring AI models are non-discriminatory and environmentally friendly, guaranteeing safety, transparency and traceability. There are concerns as to whether a global standard of AI regulation can be established due to the complexity of the relationship between AI and the environment.

For example, training AI on large language models (LLMs) can consume vast amounts of energy. However, LLMs are predicted to have a pivotal role in furthering sustainability initiatives by monitoring renewable energy consumption in smart grids (Cooper 2023). With this in mind, there may be a selection process for LLMs and similar technologies whereby companies can opt for LLMs with lower carbon footprints, with a slight cost to their efficiency, e.g. a trade-off.


There is a strong argument that those who patent AI tools, platforms and partner with companies for use of products should be subject to the same regulations, standards and ethos that other companies face to ensure good governance. This is to reduce the risk of AI being used and abused as the legality, regulations and guidance surrounding its use is as nascent as our current comprehension of AI. This could be comparable to monitoring and improving the impact of your entire supply chain, from start to end, as is advocated for by sustainability certification standards such as BCorp.


What is clear is that AI will play a pivotal role in filling our data and knowledge gaps to improve sustainable strategies and has the power to do much more. In the words of the late Stan Lee, “with great power, comes great responsibility” and it is within our best interest to ensure that AI as a technology is regulated.



Author: Charlotte Macdonald

Editors: Neil Sandy, Louis Goring-Morris



Wellers Impact is a UK-based, FCA-Regulated Impact Investment Manager which works to unlock community-focused impact through SDG-focused impact investing. Through innovative investment models that utilise fair economics, Wellers Impact originates investment opportunities across three core business activities; real estate developments in partnership with local land-owning not-for-profits in East Africa, financial support for agriculture firms and supply chains globally through sustainable development finance and direct investment into private water, sanitation and plastics recycling firms globally. Investment involves risk. Suitable for Sophisticated, Professional and High Net Worth Investors only.


bottom of page