Inspired AI Summit 2020 Top Takeaways: Episode 9
9 episodes later and we’ve reached our end! As such, we would like to thank Inspired Minds for an incredible virtual summit experience - we greatly look forward to the next, be it in-person or virtually.
The final episode of the summit, Episode 9, was titled ‘Global Game Changers’ and explored ‘AI for Good – can Tech Save Us?’. Our top takeaways focused on two topics we found particularly interesting this week: ‘Protecting forests with AI and Satellite Data’ by Indra Den Bakker (CEO & Co-Founder, Overstory) and ‘Ten years to 2030 – the global starting point for the climate countdown’ by Kim Cobb (Georgia Power Chair and Advanced Professor, Georgia Tech), Farhana Sultana (Associate Professor, Department of Geography and the Environment, Syracuse University), Nanda Piersma (Scientific Director, HvA Expertise Centre for Applied Artificial Intelligence), Tom Peacock-Nazil (Founder of Seven Clean Seas) and Adam Levy, who acted as moderator.
Protecting forests with AI and Satellite Data:
Indra Den Bakker led an insightful session on how artificial intelligence (AI) can apply data obtained from satellites to protect forests across the world.
Den Bakker opened by displaying the requirement for using real-time data in forest protection and explained the challenges we currently face as a result of climate change. Deforestation is occurring at an alarming rate, with the equivalent of 27 football pitches of forest land being cleared every minute.
The session then drew attention to how satellites can provide access to the very data required to protect forest land. Over the last decade, over 1000 more satellites have been launched - as such, there exists a greater supply of satellite data than ever before. Images provided by these satellites are now of a higher quality than was previously possible - as a result, new insights and capabilities arise from satellite data, including the ability to detect tree species and bark beetles within tree populations.
The challenge that comes with increased data usage surrounds how to apply it - this, however, is where machine learning can be leveraged to our advantage. Machine learning allows for data sets to be integrated into a deep learning model, which results in accurate predictions of high-risk forest areas and forecasts where deforestation is likely to occur. Den Bakker has used these models to create a real time vegetation platform: Overstory.
Further, Den Bakker outlined the potential that these systems hold for protecting forests. The intersection between power lines and vegetation is the single largest cause of forest fires and the use of legacy technology means that it can take weeks, or even years, to scan a whole power grid and highlight high risk areas. Using Machine learning, however, Den Bakker’s team have been able to scan entire power grids in a day, predicting risk through an analysis of location, weather and tree species. As such, these systems greatly improve operational efficiency, whilst providing customers with insights that have previously gone undetected.
Ten years to 2030 – the global starting point for the climate countdown:
The extended session entitled ‘Ten years to 2030’ acted as a deep dive into the climate countdown and explored what the global starting point in tackling the crisis should be.
Cobb and Sultana’s opening thoughts placed emphasis on the fact that current impacts of climate change, including the generational harm suffered by communities around the world, come as no surprise. Now, however, is the time to “turn science into solutions”, as Cobb put it, and address issues affecting both impoverished communities and advanced industrialised societies. Peacock-Nazil too highlighted an example of the impacts vulnerable communities face given climate change and plastic pollution. In Indonesia for example, fishing communities are extremely reliant on the ocean – warming, however, compromises the ocean’s offerings and places the livelihood of these communities at huge risk.
Addressing the debate from a technology lens, Piersma suggested that whilst the likes of data collection are important in mapping out the climate crisis, collection itself is an energy consuming activity – as such, it is vital that we rethink our usage of information technology in this fight. Importantly, Sultana also emphasised the importance of talking about ‘climate justice’ rather than climate change: given that systemic differences increase one’s exposure to effects of climate change, it is essential that underlying structural inequalities are addressed in our tackling of the climate countdown.
Further, Cobb highlighted that COVID-19 has resulted in an increase in discussions surrounding the usefulness of science acting as a tool in aid of public health. The message the pandemic has constantly reminded us of is simple: above all, we must protect public health and lives. The climate crisis similarly requires an understanding and cementing of the same message. Sultana too suggested that the pandemic has shown us what it means to ‘build back better’ - in addressing the climate countdown, now is the time to do exactly that.
Brainpool Thoughts:
At Brainpool AI, we believe that AI should be viewed as part of the solution to climate change and we are always excited to explore ways that organisations can leverage tools to assist in the fight against global issues. Den Bakker’s example of AI being a means to tackle deforestation highlights the potential that AI holds in addressing climate change. At Brainpool, we aim to unlock this potential and seek to provide organisations with access to bespoke AI that solves pressing business challenges and aids in the development of more sustainable business practices. Recently, for example, we have been engaged by Natural Resources Canada to design a platform that will forecast biomass availability and analyse bioenergy production opportunities, costs and environmental impacts under different demand scenarios.
Written by Anjali Kapila & Joe Duszynski Lewis