Climate change is getting worse by the minute and one needs to talk about it and carbon emissions. Carbon emission is the biggest source of pollution and it’s something we should keep a strict eye on. We are confronted with an ever-rising greenhouse gas problem where the blanket of clouds traps heat and increases the temperature. Currently, there is 50 GT of carbon emission annually, but where does Artificial Intelligence come in?
Statistics say that Artificial Intelligence has the potential to help reduce the emission of greenhouse gasses (mainly carbon dioxide) by 16%. Capgemini Research says by the year 2030, AI potentially might help various organizations across sectors, for instance automotive, retail, consumer goods, etc to meet 45% of the target as mentioned in the Paris Agreement.
Artificial Intelligence can help accelerate the process of transformation. AI is capable of essentially integrating and incorporating renewable energy into a stable electricity grid. Simultaneously it should be kept in mind that the AI used itself should be environmentally sustainable. All the AI models which make use of deep learning have grown in scale and breadth of their application. Researchers have realized that AI models use computing resources efficiently.
All electronics, no matter how technologically advanced they are, run on hardware. It’s no different for AI which entails necessary resources like processors, memory and networks. Computers produce two types of Co2 emissions. Electricity is what powers everything and the problem arises when the electricity isn’t carbon-free. All paramount cloud providers (like Microsoft, Google, etc) are either already powering their cloud computing data centres with 100% carbon-free energy or have roadmaps to do so by the year 2030.
The second embodied carbon emissions arise due to the manufacturing processes that create computing hardware additionally generating carbon emissions which attributes to the manufactured products. Reducing carbon emission ends up being challenging as per Microsoft’s 2021 Sustainability Report.
One needs to carry out research so AI can run more efficiently on computing hardware which in turn will consume less processor time, less memory and more. This technique is capable of reducing both the operational & embodied emissions associated with all the tasks of AI.
The challenging part of this goal is achieving the goal and simultaneously not sacrificing other desirable attributes, for instance, the accuracy of predictions delivered by the AI for which it is designed.
For more such updates, keep reading techinnews