1. AI healing our environment; Cleaning waste in our oceans
Under AI for Earth, Microsoft provides support to different organizations fighting for the betterment of Earth. One of these nonprofit organizations that Microsoft supports is called “The Ocean Cleanup” which is attempting to eliminate waste from our rivers and oceans. Initially, the organization was using security cameras mounted on their models to spot debris and identify plastic but it can be a labor-intensive process. In collaboration with Microsoft, they developed a solution that relies on AI in identifying trash and plastic from organic material. This process was replicated on cameras that were mounted on drones and ships crossing oceans.
Essentially, this AI solution has helped the organization broaden its vision and the organization aims to clean up 90% of floating ocean plastic pollution by 2040.
2. AI healing our environment; Preserving Biodiversity & Wildlife
Populations of species that are under threat can be protected as AI will provide better solutions for anti-poaching units to utilize. For example, TrailGuard AI, a camera device that detects motion using an offline, on-device AI algorithm. It is capable of identifying humans and vehicles along with alerting units in real-time. These devices can be laid out across the landscape monitoring and notifying of suspicious activity.
This innovative AI technology ultimately led to the identification of 20 poacher gangs in the initial 15 months of installation. It’s truly powerful and is helping wildlife units in protecting wildlife and preserving biodiversity.
3. AI healing our environment; Managing facilities
DeepMind, an AI-powered solutions company, offers and builds systems that help tackle a range of problems. In 2014, it was acquired by Google and as a joint effort, DeepMind was able to develop an AI solution in reducing CO2 emissions. They developed AI that teaches itself to minimize the use of energy with the goal of cooling Google’s data centers. As a result, Google has successfully been able to reduce its data center energy requirements by 35%. Machine learning has reduced the energy used for cooling Google data centers by 30 percent.
So how is this significant? DeepMind and Google Cloud are making this Cloud technology solution available globally for use. It can be implemented by airports, shopping malls, hospitals, data centers, and other commercial buildings along with industrial facilities. The ultimate goal here is to reduce CO2 emissions. DeepMind believes that “even minor improvements would provide significant energy savings and reduce CO2 emissions to help combat climate change”.
Is AI really contributing more to carbon footprint than helping it?
Well, it certainly does! But, this is not true for all models where some are more power-hungry than others. At the same time, there are ways to make artificial intelligence green. Many companies, like Google, have dedicated resources to this effort. The researchers determined that the carbon footprint of AI/ML models is heavily dependent on a range of factors. For instance, the design of the algorithm, the type of computer hardware used to train it, and others. So, altering these contributing factors can help reduce the carbon footprint of training models.