Using artificial intelligence to reduce emissions in the company

According to a survey, less than one company in 10 is able to calculate exactly how many emissions it produces. But Ai tools can help estimate impact and take action

Photo by 政 徳 吉田 from Pixabay Becoming a truly sustainable company is still a complex journey. Measurement, monitoring and adoption of solutions to reduce carbon dioxide emissions are operations that artificial intelligence can optimize in all production processes. In order to understand how much CO 2 a company emits and to reduce its emissions on average by almost 40% compared to what is currently happening.

These are the results of a survey by the consulting firm Boston Consulting Group (Bcg ), who interviewed the sustainability managers of 1,290 organizations in nine economic sectors around the world. From power companies to the automotive sector, from industrial factories to financial institutions, from technology companies to health care and insurance.

Companies do not know their emissions

There is a commitment, but the results are still poor. According to the report, 85% of the companies interviewed are concerned about reducing their greenhouse gases, but only 9% of these are able to quantify their total emissions in an exhaustive manner. The result? Just over 1 in 10 companies has reduced the CO 2 produced according to the targets set five years earlier. The problem lies primarily in the quantification of the carbon dioxide it emits. Eight out of ten companies do not include some types of internal emissions in their sustainability reports, while the vast majority do not report any external emissions. They are those related to what happens along the support chain, the company supply chain. Suffice it to say what impact delivery or agri-food companies have. In general, external emissions represent on average 90% of the overall emissions of any company.

The study shows that 87% of respondents actually intend to increase the scope of their measurements, but currently 76% it does not know how to quantify the ecological footprint of the product it makes or the service it provides. To conclude, the companies interviewed admit an error rate in their CO 2 emissions measurements that fluctuates between 30 and 40%. If they fail to measure them, it is difficult for them to establish effective plans to be able to bring them down.

Not all economic sectors are clearly at the same level: BCG identifies 4 different ones, depending on the maturity and awareness of the issue of emissions. “Stage 1”, the lowest, refers to sectors where on average there are inaccurate and complete measurements, no targets are set and therefore the emissions reductions are almost zero. “Stage 4”, on the other hand, identifies business sectors where measurements are effective, objectives are dynamic and precise and the CO 2 cut is significant.

Report BCG 2021 - Use AI to Measure Emissions- Exhaustively, Accurately, and Frequently

Artificial intelligence at the service of environmental sustainability

The AI ​​tools operate on three analyzes: descriptive, predictive and prescriptive. Three specific directions emerge from the report. As Charlotte Degot, partner of BCG and co-author of the survey describes, it is about the "identification of effective initiatives to reduce CO 2 emissions, the monitoring of results and the consequent operational optimization". Three activities that follow the significant improvement of the identification of internal and external emissions of a company and the correct measurement of the ecological footprint of a service or product.

All this, according to Degot "could lead to an average reduction of emissions produced by a company of about 40% “compared to today. Artificial intelligence increases the speed and reliability of processes. In general, the photographs of the corporate world on their emissions are static and obsolete. Suffice it to say that 86% of the companies interviewed in the BCG report record their emissions manually, using spreadsheets.

So how is AI used to better monitor and reduce company emissions? It depends on the goal. For example: to better measure the data by automating its retrieval, cleaning and matching and then extrapolating the missing ones. To set reduction targets and identify the best countermeasures, creating simulations and verification roadmaps. And yet artificial intelligence helps in managing large-scale programs, tracking and reporting in real time the variations with respect to the starting standards, performing direct and real-time optimization on processes that consume a lot of energy.

Artificial intelligence is also a source of CO 2

The panacea for all environmental ills does not lie in the digital world. CO 2 emissions, as shown by the lockdown, also arise from technological devices: this is the question of digital sustainability. For this reason, the benefits of artificial intelligence can have a high cost for the environment. "AI generates CO 2 emissions, which today are estimated at around 2-3% of the total - explains Sylvain Duranton, global leader of Bcg Gamma, the division that oversaw the research - it is possible to discuss how much they are but not the fact that they are huge and exponentially growing, as is the growth in the use of artificial intelligence. So before proposing environmental solutions through AI, you need to make sure that these are the first to be measurable as CO 2 impact ". In this regard, Bcg Gamma has created with Yoshua Bengio del Mila an open source tool called Code Carbon. It is a tool that records the amount of energy used by the underlying infrastructure of major cloud providers and data centers and estimates the amount of CO 2 emissions produced. Such as, for example, the kilometers traveled by car, the hours of TV watched and the daily energy consumed by an average family.

A tracker that can also be used to record the estimated equivalent of CO 2 produced from Ai applications used at an enterprise level. A tracking system that still needs to be perfected thanks to the input of programmers and experience of use: the theme of digital sustainability and the impact of AI is still a virgin land. We are only at the dawn of applications and therefore of evidence of the impacts of artificial intelligence in our society.


Culture - 1 minute ago

The Mirror of Water, the interactive experience to discover a submerged city


Wpp award goes to the startup that uses rice waste for construction


DeepMind's latest trick? Weather forecasts

Topics

Environment Cop26 pollution Artificial intelligence globalData.fldTopic = "Environment, Cop26, pollution, Artificial intelligence"

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.