Hyperight

How AI and Data Help Companies in Reaching the Sustainability Indicators?

desk with laptop and documents with insights and plans

A while ago, at one of the weekly AI After Work Podcast episodes, a sustainable way of living was discussed amongst other main themes. And most importantly, how the business sector is catching up with (as said in the episode) the “Green Transition”. That made us think about the role of AI and data in the transition towards sustainability of society. So, we decided to dig a bit deeper into this. Before we share some of the findings we came across, let’s first give a brief overview of sustainability, what sustainable business means, and why it is important.

In 2015, the United Nations set goals to be reached by 2030 by societies during their development, and this is better known as Agenda 2030. The main aim behind this commitment is to fulfill “the needs of present generations without compromising the needs of future generations.” Those are the well known by now the Sustainable Development Goals (SDGs). Everyone in every society has to contribute to reaching what was once agreed upon, especially the businesses, since they impact the environment through operations and the supply and value chains. A significant percentage of the worldwide greenhouse gas (GHG) emissions is from the industry. Those emissions are generated: directly in the facilities and indirectly through electricity consumption and the fossil fuels required to transport products and materials. That is why businesses need to care about sustainability. The need for that comes from employees, investors, communities, politicians and customers.

To achieve this, businesses are turning their focus and investments to data-driven solutions, technologies and ways to minimize the impact on ecosystems and societies. Based on this, the so-called Data for Good movement appeared, also known as AI for Good

A person during a business meeting
Photo by Headway on Unsplash

There is always a but. As in any process of transformation or transition, there are obstacles. And not just in the business community, but beyond it, there is a growing concern regarding the need for more significant action for reaching the indicators of SDGs.

Is there leadership inside the business to recognize the call for “green transition”? If there is, is there support for implementing strategies towards sustainability? Once that is set, is there a way to measure and report reducing the green footprint on the environment?

Understanding the Role of AI and Data for Businesses’ Sustainability Efforts

One way to define data-driven sustainability is to collect and use data to form and make decisions that lead towards measurable and sustainable business practices. Simplified would be companies to continue increasing profitability, but to use data and tech power to consume less energy, lower GHG emissions, reduce waste, and use the resources reasonably. 

To start the journey for sustainability, every business must set objectives or adjust the business goals with the SDGs, draft an action plan, and then take steps to apply AI and data-driven solutions.

The initial step involves consolidating and storing corporate data, along with external data sources, within an organization’s centralized repository.

With this step, any business organization can follow its green portfolio and decide upon investments, emissions, trends, market insights, and renewable energy data, among some of the activities. Another step in these sustainability efforts would be AI to help in managing and measuring the outputs of the business organization’s green activities. Based on the accurate insight, businesses will be able to make informed decisions if any adjustments are required to reach specific net-zero goals. A very important step in implementing the sustainability policies and/or corporate social responsibility initiatives is to have onboard other parties of the chain. AI can help here, too, by detecting and predicting anomalies from third parties of the network and, at the same time, providing a better understanding of the risks for the sustainability of a whole network.

AI and data-driven solutions can:

  • Continue to increase the automation of manufacturing and industrial practices of the businesses with the smart use of technologies such as machine-to-machine communication (M2M) and the internet of things (IoT).
  • Generate more data sets for each step along the value chain. Businesses nowadays are instrumented with sensors, trackers, and other smart devices collecting data in real-time.
  • Help businesses select and monitor their suppliers by more efficiently compiling and analyzing unstructured data such as news entries, audit reports or social media postings.
  • Purchasing companies can indirectly benefit environmental performance among suppliers with tracking of inventory and purchasing patterns for demand forecasts.
A device with insights and application
Photo by ThisisEngineering RAEng on Unsplash

Key Challenges in Reaching the Sustainability Indicators

For the 17 SDGs, there are 169 targets and 247 indicators to help measure the progress towards sustainability of societies. The Climate Policy Watcher defines indicators as “symbolic representations (e.g., numbers, symbols, graphics, colors) designed to communicate a property or trend in a complex system or entity. Traditionally, most indicators for decision-makers have been numbers calculated by statistical services, including complex indices, such as the gross national product (GNP), or percentages, such as the unemployment rate.”

Even though there was progress in achieving sustainable development during the past years, the integration of the indicators to support self-regulating sustainability is still a significant challenge.

From an analytical point of view, the management of SDGs is challenging because of the lack of data, the synergies and trade-offs in the defined SDG system. The quality of indicators is mainly determined by how reality is translated into measures and data. Most indicators have been constructed based on existing data, and generating new data takes years. This excludes relevant indicators for newly emergent issues. So, if indicators can be selected only from the existing data, the desired evolution may be blocked.

In 2021, 60 business leaders committed to the core Stakeholder Capitalism Metrics released by the International Business Council (IBC) of the World Economic Forum. These metrics are a framework that measures the long-term enterprise value creation for all stakeholders towards the environmental, social and governance (ESG) credibility. Because of the slow progress by societies towards targets, at the 2021 G7 Summit, all attending parties committed to a 2050 deadline to achieve net-zero emissions. The commitments from the policies will reflect on the business sector, and that’s their so-called compliance risk. Because many do not know how they are doing against their targets on a day-by-day basis. 

Interestingly enough, according to the Edelman Trust Barometer for 2022, there is less and less trust in the business’ societal role, including climate change (52%), economic inequality (49%), workforce reskilling (46%) and trustworthy information (42%). Among the key findings of the same report, is the expected societal leadership as a core function of businesses. CEOs are expected to shape conversation and policy on jobs and the economy (76%), wage inequality (73%), technology and automation (74%) and global warming and climate change (68%). The challenge is for the business sector to respond to this call and play a central role in a sustainable future. This is the reputational risk that businesses have in their mission for sustainability. 

Businesses face a strategic risk, too. One starting point to cope with that challenge is the transparency and open reporting of the activities. Another strategy is increased collaboration with the civil and public sector, peers and competitors towards viable and scalable solutions. 

Adequate progress requires the processing of massive amounts of diverse data to provide immediate insight that operational managers can act on in real-time to hit targets – and the only way of doing this is to leverage digital technology. But does this require extra investments? Is there deep industry expertise? Businesses face the operational and financial risks. Any businesses that fail to find the answers and adjust accordingly will risk damaging their reputation shareholder value and incurring consumer boycotts. They reduce their competitiveness and operational efficiency and burden themselves with unnecessary costs, making it harder to attract investment: 95% of fund managers see oil companies not responding to climate change as unsuitable investments. 

At the same time, we should not ignore, and we should consider, some of the negative impacts businesses have on the environment with AI /ML solutions. For example, one of those effects of AI results from hardware and infrastructure use, like data centers and networks. This is causing increased consumption of resources and energy. According to the EU findings, data traffic in digital networks and other information and communications technologies (ICT) consumes about 7% of the world’s electricity today. It has been estimated that 5-9% of the world’s total electricity use is caused by ICT, rising to 20% in 2030. Greenhouse gas emissions (GHG) from global data centers and communications networks are estimated at 1.1 to 1.3 Gt CO2eq in 2020. One example on how to tackle this challenge, is the Montreal Institute for Learning Algorithms. They released a tool designed to estimate how much carbon is produced in training machine learning models.

Learning From Best Practices on How to Use Data to Model Sustainability

Suppose we can agree that the sustainability model combines brand image, cost savings, regulatory compliance, and competitive advantage through innovation. In that case, we also must agree that there are positive examples of businesses using data to shape that model.

Two women testing a AI and ML model in a working station
Photo by ThisisEngineering RAEng on Unsplash

A typical case of those positive practices is IBM. The company (one of those 60 committed to Stakeholder Capitalism Metrics) acquired Envizi, a data and analytics software provider, for its environmental performance management. IBM plans to integrate Envizi with its existing solutions to bring together the everyday operations, sustainability data and strategy to move faster in achieving their goals.

Another example is AVG Group Sarl. This investment firm in Norway engaged the software platform Rio AI to better understand their ESG performance to decide which areas to improve. The final report of the assessment is a summary that showcases the positive environmental impact and visualizations of environmental data. It demonstrates the direct impact of individual investors through crucial metrics such as carbon emissions saved, homes powered, and equivalent trees planted.

In 2019, Shell and GHGSat signed a framework agreement according to which GHGSat satellite-based monitoring services will obtain methane (CH4) emissions data of certain agreed Shell facilities globally. Chevron and TotalEnergies supported the research project to achieve a world-first in demonstrating high-resolution satellite-based monitoring of CH4 emissions at sea.

Here are some other more broader examples of how deployment of AI/ML initiatives can have positive impact of the environment: 

For one of the significant problems, the air quality, and using the insight to make more informed decisions, Google has equipped several of our Street View vehicles with air pollution sensors to measure street-by-street air quality on city streets. Copenhagen, Amsterdam, London, Dublin and others are just some cities that use Air Quality maps. 

FrostStrat project is another case where we can see deployed tech tools that support farmers regarding forecasts to carry out anti-frost measures in time. The company’s software helps in climate data recording, the creation of weather forecasts, and the alerting of farmers.

Machine learning to help save the bees? A hive monitor using an open-source ML framework is helping in collecting data. A camera records things like how many bees return to the hive every day, how they’re moving, and if they’re carrying pollen. Afterwards, the data is shared with experts to make more informed decisions on where to plant trees and flowers, for example.

The Norwegian Oil and Gas Association uses AI and 3D printing to lower expenses and increase the flexibility of the oil and gas company in Norway. Similar user cases are constantly growing in the manufacturing industry, consumer goods, logistics, procurement, etc. We can agree on the conclusion with those researchers that say that no single company can achieve sustainability on their own and that the key is the partnerships among the organizations. And that solutions for overcoming data gaps in SDGs can be applied globally.

Conclusion

Whether it is the targeted year of 2030 or 2050, the time to act towards sustainability is now. Individuals and organizations must take specific speedier steps for global social, economic, and environmental development in the upcoming years.

The issue of sustainability is becoming critical for businesses, and based on that, AI and data-driven initiatives will become a substantial percentage of any organization’s investment in technology. We say critical since the business sector faces: strategic, operational, financial and reputational risks if not moving towards (profitable) sustainability.

Even though progress has been made, there is a need for further development to tackle, measure and report the achievements. There isn’t a simple answer to how AI and data can help reach sustainability indicators. But it’s simple to take small steps in driving sustainability through culture and people, adjusting and digitizing operations and maintenance, then applying initiatives in the supply chain and engaging in supporting the societal challenges.


Featured image credit: ThisisEngineering RAEng on Unsplash

Add comment

Upcoming Events