Whether one likes it or not.
Whether one is familiar or not.
AI has gained a massive footprint in not only the Technological sector but also across Education, Research, Marketing, Content Writing and many more.
Use of basic to more advanced and sophisticated AI tools & techniques is much required and often giving competitive advantage.
Not only this, big tech and other companies are announcing Million to Billion Dollar investments in AI Data Centres and other initiatives.
However, all this usage and expansion comes at a cost.
AI Data Centres run on Electricity and leave a big carbon footprint.
Electricity and cooling needs if not coming from renewable sources sap up a lot of environmental resources and threaten to derail Net carbon neutrality (or carbon footprint reduction) aims to a big extent.
While companies like Google, OpenAI are trying to come up with solutions like Data Centres in low carbon footprint areas, or colder regions or going for Liquid immersive cooling, we as users can also help in reducing the environmental impact by these means:
1)Club together multiple prompts. Make a list of points - Bullet or numbered - from 1 to N and try to ask in one or least go.
Repeated prompts means everytime the backend engine of AI uses a lot of electricity and compute resources. One go reduces this.
Use a Prompt Library if needed.
2)If it is a very simple thing you need to ask or know, just search on search engines like Google, Startpage, Bing, DuckDuckgo etc.
They use much lesser energy than AI tools or Apps/Sites.
Or in short, use only as per need and "if needed".
3)Do not build a whole ecosystem or a completly new Bot or even LLM/SLM if an existing available system fulfills your need.
For e.g- Custom Gems on Gemini or Custom GPTs on OpenAI-ChatGPT can be used as a base to develop systems that you want to build like a customized Ticketing Site or a Job Search utility or even some kind of an Ecommerce system.
This would save immense resources.
4)Similar to point 2 above, use systems that fulfill your need. For e.g- for simple code debuuging use ChatGPT/Gemini/Claude simple or basic models.
Do not use Reasoning or Pro models.
Use Pro or Reasoning Models for preparing Scientific case studies, help in writing Thesis etc.
Or try using a local LLM in some cases that would consume even lesser resources.
I have not gone into the numbers as that could perhaps be a separate article altogether.
Yet based on some sessions I have attended and what I have read, above and more points guide us on more optimum use of AI resources and help keep carbon footprint and overall environmental impact in check.
For more such thought provoking Articles, insights and discussions, do come and join the wonderful world of ForumCoin. An English language forum, where you can discuss and share ideas and opinions on a variety of topics.
For sign up, click here :
https://forumcoin.com/index.php?r=5080
