Artificial Intelligence is a rapidly evolving technology with new use cases popping up on an almost daily occurrence. AI’s foundational large language models are praised for the ability to generate output based on huge datasets and human refinement. Now, thanks to a partnership between IT service provider CGI and geospatial company Ordnance Survey, those models are being used as predictive engines to help clean up the UK’s water pollution problem.
The first trial of the joint project saw the AI-powered tool predict pollution events (based on historical data) with over 90% accuracy. It works by providing the AI with a bunch of information it then analyses, learning patterns that can be used to foresee future incidents. The team trained the model on the combinations of environmental events and industrial activities. For example, looking at situations where agricultural chemicals were washed downstream due to large rainfall events. In future, the team hopes that the AI will be able to advise surrounding businesses as to how to avoid becoming the catalyst for water pollution.
We’ll give it all of the geographic information, as well as data sets from the sensors for it to learn and develop the predictive mechanisms to be able to inform where these incidents are occurring and indeed when they will take place.
Mattie Yeta, CGI Chief Sustainability Officer
Whilst the trial run was confined to the North Devon Biosphere Reserve, a 55 square mile/142 square kilometer area, future plans would see the tool rolled out to locations across the UK.
The team are optimistic that the project could be initially used to help address the poor water quality at Combe Martin, a nearby seaside town. Residents are used to being advised not to swim due to algae, sewage and agricultural discharges that are commonly present in the water. A tool like this could help in these efforts and restore confidence in the tourism industry, a staple of the town’s economy.
Hit the link to find out more, including how real-time data, including oxygen saturation, ammonia levels and pH balance, is collected by floating water sensors and relayed to the AI via a network of Wi-Fi networks.
