Of Greenwashing to a real corporate social responsibility (CSR) policy: where are we really? Since the establishment of the European Directive on the publication of extra-financial data (CSRD), companies with more than 500 employees must analyze the financial risks to which they are exposed but also the environmental consequences of their activities (quantity of CO2 emitted , damage to natural environments, etc.). This sustainability report will reflect the reality of their CSR policy. However, in the application of this European directive, France is rather ahead. Little by little, a whole movement is developing, with on one side the constraints of the law, and on the other, the weight of employees convinced of the social and environmental role of companies. Even new technologies in general and AI in particular are putting themselves at its service. Silent, the CSR revolution is inevitable.
It is called Barbatus, in reference to the bearded vulture, a bird of prey threatened in France. Launched by the start-up Iceberg Data Lab in the summer of 2023, this conversational artificial intelligence should assist sustainable finance analysts specializing in Environment, Social and Governance. This prototype – a derivative of OpenAI’s ChatGPT – is supposed to allow these experts to more easily find the right data in the thicket of information published by companies, such as their financial reports. A way to help them “save precious time by being a compass for directing investments towards greener industries”, underlines Pierre-Olivier Haye, the technical director of Iceberg Data Lab.
Like this fintech nugget, which has just raised ten million euros, technology specialists see many possible applications for AI in the corporate social responsibility (CSR) sector. This type of software with advanced data analysis and management capabilities is indeed welcome in this complex because cross-cutting theme. They could be used in carbon footprint analysis, supply chain monitoring or to calculate the social impact of company operations. “The objective is to progress on the reliability of data because it gives credibility to the actions of progress”, underlines Fabrice Bonnifet, the president of the College of Directors of Sustainable Development, an association which brings together 350 executives in the sector mainly from large companies based In France. AI-based solutions should help CSR experts better “respond to the multiple heterogeneous obligations of regulatory and voluntary reporting”, he adds, for example by publishing indicators on an organization’s water consumption or the volume of waste produced. Beyond CSR as such, AI finally promises better optimization of all the resources used by an organization.
Not yet a priority for businesses
Although these tools already raise significant expectations, many specialists remain in suspense. “In the daily life of businesses, it is not yet the priority, warns Fabrice Bonnifet. We are starting to talk about it because AI will become essential. But we also clearly see the criticisms and the associated limits, in particular it will be necessary adopt an ethical reflection on uses.” Generative artificial intelligence raises questions. They train on gigantic masses of data. An approach that seems not very virtuous compared to some of the uses that are made of it. “Smaller language models should emerge as the most sustainable,” agrees Pierre-Olivier Haye.
His company is working on directly hosting its conversational AI, Barbatus, so that it is more frugal. Knowing that behind the fashion around generative artificial intelligence, there are more classic uses of AI which already work well. At EcoAct, a decarbonization consulting company, a subsidiary of Schneider Electric, data scientists use machine learning algorithms to assess the carbon footprint of organizations through hundreds of thousands of lines of information. “We are not trying to build black boxes, but explainable models to have the most transparent approach possible,” summarizes Nadège LEspagne, director of Climate Data Analytics. A way of also showing that these tools are compatible with a virtuous approach.
.