Companies use AI for their ecological transition
In the aftermath of the floods in southern France that killed 11 people and seriously damaged the Aude department, Prime Minister Édouard Philippe announced his intention to implement an “accelerated natural disaster procedure”. But in practice, how can we manage better the risk of natural disasters? These are likely to increase with climate change, but luckily AI can be the solution to many environmental problems. MP Cédric Villani expressed his desire to develop artificial intelligence solutions to fight climate change and address the necessity of an ecological transition by unveiling a report on the development of AI on March 29. Among the topics discussed were the identification and preservation of biodiversity, damage repair, and the modelling of the impact of mankind on the environment using AI. Progress on the latter theme would help prevent future weather-related disasters. Indeed, the devastation caused by the floods in the Aude was partly due to excessive urbanisation with over-concrete grounds which were thus poorly waterproofed. Visual recognition, thanks to the analysis of satellite images, is already fully capable of mapping the human hold on the territory through regular rate analyses of concrete, and can thus automatically detect a danger.
There are many ways that AI can be mobilized to help people understand their impact on the environment and act accordingly. In France, we can consider creating a database including electricity consumption (via the Linky meter), air pollution data, rainfall and sunshine. The use of these figures would make it possible to adjust needs and therefore production in real time, and could reduce polluting emissions. We could also map living species or deforestation in order to restore regional and global ecosystems.
These observation tools then make it possible to act effectively and accurately to protect the environment. And it is this precision of analysis that makes the strength of AI, especially in sensitive areas such as coral reefs. Indeed, it is extremely difficult to document and therefore protect underwater life, but the development of intelligent robots equipped with visual recognition technologies makes this possible. For example, the RangerBot developed by the University of Queensland in Australia is not only able to monitor the health of the Great Barrier Reef, but also to accurately identify (99.4%) the starfish that decimate the Reef and inject a deadly toxin into them. The robot uses a computer vision system to move and monitor, instead of the usual acoustic systems. This technology is not only more accurate and relevant, but it is also less expensive to produce, making it possible to consider large-scale deployment.
The question of scale is at the heart of the challenge of an ecological transition. That’s why Veolia Recyclage called on Max AI at its Amiens site in France, an intelligent robot that automatically sorts household waste and improves its recycling. It is able to perform 3600 sorting operations per hour, compared to about 2200 for a human operator, and to recognize certain types of unwanted waste on the conveyor belt of a sorting line and remove them from it. As France has set itself the objective of halving the volume of waste sent to landfill by 2025 and recycling 100% of plastics by then, it is essential to sort better to improve recycling, and all the major players in the sector agree that AI is the solution. Veolia is not the only group to modernize its sorting centres. Suez has invested 15 million euros in its centre in Val-de-Marne, equipped with six optical sorting machines, and Paprec commissioned a centre last year equipped with optical sorters in the Alpes-Maritimes.
The Climate Marches in France, the victory of the Green Party in Bavaria and the progress of environmentalists in Belgium recently demonstrate that citizens are becoming aware of the ecological urgency. On the associative side, the AI for Good initiative launched in the wake of the Villani report aims to bring together the various actors (associations, AI researchers and entrepreneurs, etc.) to find solutions “to major societal issues”. To encourage companies to participate in the ecological transition, the MP recommends in particular the creation of specific funds, intended to finance, via the BPI or the FrenchTech, companies carrying out environmental projects. New advances in AI can therefore drive new initiatives by companies for their own good and that of the planet.