Updated: Jan 24
Sustainability has been a long-standing concern in our economy. Standing on the tipping point of an environmental crisis, promoting a sustainable economy is one of the key focuses of government level policies, industry level revolutions and business level transitions. In light of the COVID-19 crisis, the UN has urged countries to become trailblazers for “green recovery”:
“With this restart, a window of hope and opportunity opens… an opportunity for nations to green their recovery packages and shape the 21st century economy in ways that are clean, green, healthy, safe and more resilient”
— UN Climate Chief
In today’s context, when evaluating how we can better promote a circular economy, we need to first appreciate the fact that most households and businesses are already aware of the responsibilities they have in leading development in a more sustainable manner. With much higher public awareness, it then comes down to the problems of practical implementation and how to incorporate sustainable methods to make concrete changes. This is where the the model of circular economy could come into play. It provides us with guidelines and principles that we could base our ambitious transitions on.
By leveraging today’s fast developing AI technology, the transition to a more mature model of circular economy can be accelerated for each stage of the product cycle— and AI has a fundamental role to play. This article will provide an introduction to the circular economy model, identify how AI can be incorporated to each stage, and discuss different case scenarios that have been implemented. This article currently focuses on the businesses' (supply side) potential, but in the future, there will be a series that highlights the consumers' (demand side) and government regulators’ perspectives.
What is a circular economy?
According to the Ellen MacArthur Foundation, the three fundamental principles of a Circular Economy are to:
Design ways to eliminate waste and pollution
Keep products and materials in use
Regenerate natural systems
Many have probably heard of many different definitions of circular economy— it can get confusing. But generally, we are seeking for a systematic and scientific method to shift from a traditional “take-make-waste” one way flow of resources to a circular based approach. It is not simply about recycling, but also rebuilding an environment that we have disrupted, as well as trying to design a new system such that the externalities we exert on the environment as we produce and consume are mitigated. The resources used are being reused, recycled, re-manufactured by a closed loop process. This chart created by the Ellen MacArthur Foundation gives an overview of what we might envision for a circular economy:
Adopting and transitioning to the circular economy is a paradigm shift for all players in the economy: government policymakers, businesses and individual households. Recognizing that the transition requires the “audacious” use of innovative technology would be a key part to accelerate the process. I came up with the following roadmap, which gives an overview of the potential for Artificial Intelligence to be utilized in each of the stages identified by the Ellen MacArthur Foundation— from raw material extraction, manufacturing, to distribution, customers, and waste handling. Nevertheless, an efficient transition to close the loop requires every party to be open-minded and flexible towards adopting new technologies, as the cooperation between multiple stakeholders is a crucial factor to realize the value of AI in every component.
The business potential of AI in promoting a circular economy
1. Environmental impact analysis and monitoring
From CSR to ESG, most businesses today have good intentions to stand up to their environmental and social responsibilities. Nevertheless, in order for them to reduce their environmental impact, they need to first estimate the effect of their activities in quantitative means. Producing such estimates could be challenging, especially for large businesses involved in multiple chains of production and distribution. This could be resolved by the use of the IOT ( Internet of Things) and smart sensors. The IOT allows the automation of accurate and reliable data collection, which can then be fed into data analytic algorithms that produces output analysis of the different sources of externalities or pollution. Businesses can thus gain insights from these analyses and take action. For example, by identifying the greatest source of residuals or pollutants in a manufacturing procedure, businesses could invest in developing or switching to more efficient production methods to cut their impact.
2. Big data powered product innovation
By using big data, product development can be done in a more evidence based manner. Two of the main product features that would be desirable for circular economy are modularity and durability. The benefits of greater durability are obvious; on the other hand, modularity means to decompose a complex product engineering process into simple subparts. Greater modularity could then make product remanufacturing and recycling more convenient, it “permit[s] the arrangement of components in a manner that can be easily modified, enhanced, exchanged, or proliferated” (Tucker J. Marion, 2010). During the process of product development, data can be collected from prototyping and testing the product. This data can now be analyzed iteratively through machine learning algorithms, which assists with the evaluation of these desirable sustainable product features, and could be used to improve upon the current design of products.
3. Blockchain and cryptographic anchoring for supply chain management
Another challenge in tracking and recognizing environmental impact of businesses is following and tracing back to the source of its inputs. What blockchain could do is to make the whole supply chain “transparent,” where each stage in the chain is recorded in an immutable way. Imagine you are operating a small restaurant to provide tasty dishes to customers; you would need to purchase the ingredients first. To make it simpler, let’s focus on the meat supplier. In order to know whether or not meat supplier is operating in an ethical and sustainable production process, you need to know which livestock farm they got their “inputs” from. If you would like to dig deeper, you might also want to know the source of the food they fed to the livestock. This could quickly become complex and difficult to track, and it is indeed a general concern of many businesses, especially for ones that have to manage a large variety of suppliers, such as Walmart. Indeed, Walmart was one of the earliest adopters to test the application of supply chain management by using blockchain to trace pork in China, authenticate transactions and facilitate accurate and efficient record keeping.
4. Smart inventory management
Smart inventory management mainly concerns the accurate prediction of customer demand to efficiently produce the right amount of product at the right time. Stockpiling could be extremely wasteful and costly for a business, especially in cases where the product cannot be stored for a long period of time— this is either because of the nature of the product (fresh food), or because the product has decreased in value over time (fashion products). By using internal data— such as records of past sales, customer preferences— and external data— such as competitor’s performances, market demand fluctuations and patterns— AI’s prediction capabilities can be utilized to prevent stockpiling and excess inventories. This not only reduces the inventory rental cost for companies, but also greatly lowers the amount of waste and unused products that may impede for transition to a circular economy. There is already wide usage of smart inventory management using Internet of things and machine learning; an example of such a service provider is “Zenventory”.
5. Automated Optimizing Delivery and Shipping
The uses of AI have enabled us to improve the logistics of shipping and delivery by designing the fastest route. On one hand, it could analyze customer order data to make best plan for shipping in different regions; on the other hand, real-time traffic data can be used to produce efficient scheduling of deliveries. Moreover, the scope of AI usage in delivery extends beyond backend planning to autonomous shipping. In 2018, Rolls Royce and Finferries launched the first fully autonomous car ferry. The cost advantages of using the autonomous truck in B2C (business to customer) last mile delivery is substantial, with the potential of reducing delivery costs by 10% in comparison to traditional delivery method (McKinsey, 2018). The design of the fastest route reduced the amount of pollutants created by shipping vehicles, especially for overseas shipping, while autonomous delivery acted as a solution to effective, low manual input deliver, allowing more funds and human resources to be devoted to more productive usage.
I hope this might give you some inspiration in how to kickstart your business’ use of AI to facilitate our transition to a brand new circular economy. Stay tuned for upcoming articles in this series on the use of Artificial Intelligence in consumer and government scenarios.