In just 40 days, Chat GPT reached 10 million daily users. The world has been amazed by an artificial intelligence (AI) chatbot that seems to know everything you ask of it. Behind their answers was a long process of training language models that show two sides of the same coin regarding environmental impact: on the one hand, The amount of solutions it can provide in this matter and, on the other hand, the high energy consumption required to develop these processes..
It has always been said that AI is here to be a strategic ally for sustainable development.. And there is no doubt that it is. It is a tool that will help us better design, implement, advise and plan the future of our planet. Technology that will allow us to build more efficiently, use resources sustainably and reduce and better manage waste that we produce, among many others.
The five laws of a sustainable agenda that demand evidence
The Argentine biologist who has rescued more than 400 Andean condors
but, The race to develop increasingly powerful digital brains is not without stakes and controversy. about its impact on our lives. Automation and staffing, bias from algorithms, reliability of information, privacy, asymmetry of power between the actors involved, and many others – which even led to a statement by experts and company executives, among others. Development- part of the general conversation in recent weeks. To these issues – very important and which need to be addressed – we must add the environmental aspect.
A study in the United States reported that in the first phase of Chat GPT training, 1287 MWh were used.which is equivalent to 550 round trips between New York and San Francisco, cities that are about 4,500 kilometers apart. Artificial intelligence uses more power than other types of tools A single model’s training process can require more electricity than 100 high-net-worth families use it in a year.
The growth of these technologies, more in the race in which the technology giants seek to generate the most powerful tool, requires processing power, as well as efficient storage and search, which means a significant increase in energy and cooling resources that require large processing centers. We must add to this lack of transparency and even the difficulty that exists in understanding the real consumption of applications and, therefore, their carbon footprint.
The big question that arises is whether We have the infrastructure, power supply and data centers needed to meet this challenge in a sustainable way. In other words, if we can ensure that technological development is in line with the new and necessary standards of sustainability.
The same thing has happened with different industries, but now with industries standing as a flag of the future. in this sense, Reducing the carbon footprint or compensating for training will be a new battle for AI.
Far from being a “stone in the shoe”, rethinking production methods from a sustainable perspective is one of the necessary aspects of ensuring long-term success.
On the other side of the coin, the possibilities that AI opens up when it comes to providing solutions with environmental impact are enormous. For example, systems that make energy consumption more efficient reach almost all industries and can be applied in different ways. There are also sensors to reduce water consumption, more specific weather forecasts that save resources in the field, smart buildings, and trash bins that get recycled, among many other sustainable solutions.
Some companies have already faced the challenge of producing “sustainable technology”. American company Dale in the past nine years managed to reduce the power consumed by its servers by 83%. Can similar goals be set in AI production patterns? Googlewhich also has its own developments in this new technology, Proposed to operate carbon-neutral 365 days a year from 2030.
It also seeks to provide tools to its clients in this field. in 2022, The company launched active help, which uses data and machine learning to perform various tasks, including measuring your carbon footprint. Using artificial intelligence, it identifies neglected projects within the company and their impact on the environment. After calculating its entire portfolio, Google discovered that more than 600,000 gross kilograms of carbon dioxide were likely idle: getting rid of it would mean a similar benefit to planting 10,000 trees. from argentina, NODS Corporation, a company providing services globally, Evolution of digital solutions crossed by sustainabilitySeeks to mitigate its impact on the basis of efficiency by design.
To accompany this whole process, it will be necessary to link renewable energies to the energy supplies of the processing centers. Investing in clean energy and using it to keep pace with the process of technology growth, especially artificial intelligence, constitutes the path that must be followed to sustain this process.
Green economy: Cabify and Uber are investing thousands of euros to promote sustainable mobility in Spain
Those who look at only one side of the coin and talk about the high energy consumption of training AI will be right. But it is also true that the number of solutions offered by this technological development is sufficient argument to see which side of the balance tends to the future. Just as regulations are necessary to guide this growth while taking into account various social and even political aspects, they are also necessary to address environmental risks.
Mitigating the environmental impact of technology is also an unavoidable requirement for the sustainable future we need.