Artificial Intelligence (AI) in Sustainability

In continuation to my earlier contribution to ‘Sustainable Design’, Sustainable product design has proven to be critical in attempting solution to such problems based on entire life-cycle (creation to disposal).  Artificial Intelligence (AI) seems to revolutionize many industries rendering them more efficient, effective and sustainable.

AI is simply a computer system that performs tasks that would invite human intelligence – visual perception, speech recognition, decision-making and language translation.  Machine learning (ML) – Analyzes large amount of data as for example, decomposition rate of compostable waste, deep learning- can be used to identify waste items in a landfill and computer vision – monitor water quality in waste water treatment plants, are a few types of AI which can take care of composting and wastewater management.  Understandably, the role of AI brings around improved efficiency, quick decision making, increased sustainability and more importantly better waste management.

GREEN INNOVATION:

The fact that everyone is concentrating on the vital options towards carbon neutrality and zero emission facilitated many novel suggestions and Artificial Intelligence (AI) seems to be one of the options at least for mid-to-low-income countries.  Understandably, any attempt towards green innovations bank on building new technologies based on existing strengths.

It is believed that AI has the potential to transform business practices and industries in addressing major societal issues including sustainability.  The attempts so for, on the crucial issues like degradation of natural environment and climate change presumably have had limitations as we still have not found an appropriate solution.  It is in this regard that we need to look at AI as it is believed to support derivation of organizational processes and individual practices to reduce natural resources and energy intensity.  The question is not how AI would address reduction in energy, water and land use but, more importantly how it facilitates and fosters environmental governance.  It is suggested that the future studies of AI for sustainability should encompass multilevel views; systems dynamics approaches; design thinking; psychological and sociological considerations and more importantly, economic value considerations without introducing long term threats to environmental sustainability.

The world continues to struggle in transforming human activities towards long-term survival even after more than three decades after Brundtland Commission Report.  The challenges understandably are quite complex demanding technological expertise towards immediate and long-term solutions.  It is in this regard that AI sounds promising where, machines can learn from experience, accommodate new inputs and perform tasks.

There seems to be three major advantages, one – automation of important, repetitive and time consuming tasks (allows humans to focus on higher-value work); second – allows insights into massive unstructured data generated by videos, photos, reports, business documents, social media posts, or even e-mail messages; third – can integrate thousands of computers and other resources in solving complex problems.  What one is interested is how AI facilitates and fosters effective environmental governance.  Decision making process banks on how formal and informal rules govern humans in decision making ability combined with how the society determines and acts on goals and priorities in managing Natural Resources.  Competing social values however, make environmental governance controversial – long-term effectiveness depends on narrowing the gap between science and policy.   The two principle barriers – information asymmetries and bias of human emotions can be overcome through AI.  In going beyond the current thinking patterns to devise science-based solutions and policies attract a paradigm shift.

AI is a facilitator offering humans to think, plan and execute holistic solutions to environmental degradation / climate crisis.  Although humans create originating architecture of AI applications, the resulting decisions will be different from expert humans. The merit of AI success depends on how well they navigate and influence psychological/sociological and organizational factors that impede human progress.

ACCELERATING SUSTAINABILITY THROUGH AI

It depends on what industry we are considering for AI application.  In general, there are the following 5 approaches which are vital:

EXAMPLES:

COMPOSTING:

Temperature and humidity being very important, AI helps in monitoring and controlling them to the most ideal conditions for waste decomposition at optimal rate.  They are also useful in waste identifying and classifying which makes it easier to target volume and composition

A city in the United States has been successful monitoring temperature, humidity and other factors through machine learning.  They are indeed the vital parameters in composting.  Predicting decomposition rate and compostable waste further improves efficient composting

WATER PURIFICATION AND RECYCLING TECHNOLOGIES:

AI while helping quality and pollutants through monitoring allows more effective water purification and recycling.  It further helps optimize treatment processes, reduces amount of resources for water treatment and improves overall sustainability.  Since it helps predicting problems in waste water treatment plants, improved efficiency and sustainability are in a way guaranteed.  Equipment failure allows proactive maintenance and reduces system failure.  Energy consumption gets reduced leading towards sustainability.

A large manufacturing company uses machine learning to identify and classify waste water contaminants allowing more effective treatment.  It also predicts volume and composition of waste water allowing more effective waste management planning.

AI powered water Recycling system is possible even in an agricultural company where large amounts of waste water is generated.  They can even be extended under predictive maintenance exercise.  This application helps identifying equipment failure, allows proactive maintenance and reduces system failures.  Optimizing treatment processes reduces energy consumption and improving sustainability as well.

AI AND CLIMATE CHANGE:

Having understood the role of AI in a few examples, let us now turn to climate change and explore the effectiveness of AI.  It is believed that AI can predict weather, tract icebergs and even identify pollution.  There is no doubt that huge amounts of data can be processed to facilitate decisions in transforming industries.  Considering that almost four billion live in areas highly vulnerable to climate change and that around 250,000 extra deaths annually is predicted between 2030 & 2050 from malnutrition, diarrhea and heat stress, the subject deserves greater attention now.

A few examples where AI could be useful,  have been depicted in the figure 2 above.

There are almost nine different areas where AI could be useful as brought out in the figure. It practically covers many of the aspects that are relevant to climate change and a combined effort may perhaps provide the direction that we all have been looking forward to.   If I take the first one on ‘Iceberg melting with the rise in temperature as has been recorded in recent times, I am afraid we may lose some of the smaller islands with the rise in sea level.  Not that we are unaware of the current consequences but, looking at all of them holistically may provide a significant insight.  The amount of data that needs to be handled is indeed very huge and AI comes pretty handy in this regard.

LIMITATIONS OF AI:

Since the application warrants large amounts of quality data, getting accurate data could sometime be impossible.   It further demands special expertise and resources. Reliability and accuracy of AI systems where algorithms used are not understood would make it redundant.

A certain amount of ethics is necessary while processing the data related to waste management especially when personal data is processed.  Increased consumption of resources for data processing and energy consumption would raise concerns on its contribution to environmental degradation.  Human oversight is another factor that would render AI application’s inability to deliver appropriate results.  It is here that human judgment scores over AI systems in decision making.  One has to be therefore careful in selecting unique needs and challenges of different waste management scenarios.  It is thus warned to proceed with AI applications with utmost care to ensure that they are designed and implemented in a responsible and sustainable manner. 

While there are such limitations, sustainable product design looks bright moving towards innovative and impactful applications in many aspects including sustainability.

CONCLUSION: