The Science of Where Magazine meets Abhishek Gupta, Founder and Principal Researcher, Montreal AI Ethics Institute.
– Looking at artificial intelligence, as you point out in your reflection The machine’s rage against the planet, it is necessary to consider the environmental risk that derives from a “heavy” design process. Why does artificial intelligence risk “eating” the planet?
AI risks eating the planet because it is trending towards being embedded in all aspects of our lives. This means that it is replacing and displacing a lot of existing processes and thus directly and indirectly reshaping our planet in a structure that is more amenable to feeding the need for gathering more data and moving it around so that we can continue to build and power such systems.
– Can you explain Green Software Engineering to our readers?
The principles of GSE are geared towards designing, developing, and deploying software systems so that they are more energy-efficient, hardware-efficient, and carbon-aware, all in the service of reducing the environmental impacts of these systems. GSE is a guide for all stakeholders in the software development lifecycle to carefully consider their choices so that the overall result from the system and its use is greener.
– As a Magazine we deal with the culture of technology. Following your reasoning, it would be considered that Green Software Engineering could make a great contribution not only to the post-pandemic phase but, above all, to the achievement of the United Nations SDGs 2030. What do you think ?
Technology is a tremendous lever in helping us achieve the UN SDGs but in the process technology itself can leave an environmental impact. More so, during and post-pandemic, our utilization of technology to undergird our everyday activities has only intensified as has the associated environmental impact of technology through higher resource usage. Keeping that in mind, we must be cognizant of how we use technology to address the SDGs and ensure that as we use it to address the SDGs, we do such that the technology itself has minimized environmental impacts.
– Artificial intelligence is a fundamental part of “the science of where”, a complex of technological solutions that deal – in various fields – to geolocalize, organize and manage the data we produce every day. What do you think of data-driven development? Will it be more sustainable?
There is a lot of progress in the field of Green AI and Sustainable AI and solutions are being generated and proposed, such as in “The Imperative for Sustainable AI” (https://thegradient.pub/sustainable-ai/), that outline concrete steps that all the stakeholders involved in the AI lifecycle can take to build more environmentally-friendly AI systems. In particular, data-driven development that focuses on profiling the various parts of the AI system and tracking their impacts through the lifecycle will help us make more informed decisions when coming up with more sustainable solutions.