AI development in the social economy: technology at the service of social impact
Artificial intelligence (AI) is booming. With the arrival of ChatGPT, a hundred experts ask in an open letter dated March 28, 2023, to pause the development of artificial intelligence in order to take time to consider its ethical aspects. Our previous article has shown that in technological sectors in tension such as artificial intelligence, the non profit nature of organizations is no longer sufficient to ensure the values defended by the latter. It is partly for these reasons that social economy organizations are generally wary of the development of these new technologies. Yet as we shall see, the development of AI in the social economy could put technology at the service of social impact.
Before looking at their uses, it is relevant to recall what AIs are and how they work. Strictly speaking, an AI is a set of techniques that allow machines to simulate certain characteristics of human intelligence. In the case of Chat GPT, the software bases its operation on a neural system called GPT Generative Pre-Trained Transformer. This is a machine learning model that analyzes and decodes the input text to provide a response to the user based on a large corpus of data. The software generates responses based on statistical patterns and word associations rather than actual understanding of the content. Although it may resemble this in some areas and with some specific training, experts are not unanimous in describing the software currently in circulation as artificial intelligence with reasoning capability.
Through this article, we propose to explore the possibilities and uses in the social economy of these software developed with learning technologies and which tend to approach an artificial intelligence.
Artificial intelligence to improve targeting of donation campaigns
Organizations have already seized this opportunity to develop tools that use artificial intelligence to improve the effectiveness of fundraising campaigns. Among their customers are big names like the NGO Greenpeace which used this kind of service during one of its campaigns to improve its ROI. As the graph below proves, the results are indeed quite convincine.
Red is the classic non-targeted canvassing, green is the classic targeting and blue is the targeting done with algorithms. Artificial intelligence improves the efficiency of donation campaigns by identifying donor profiles and directing mailings in the best possible way. It is therefore a gain in efficiency, profitability but also an ecological argument by reducing the number of emails and therefore the carbon impact of a campaign.
Artificial intelligence would therefore allow to reach more efficiently certain contributors. However, there are some reservations about the large-scale deployment of the tool and its real effectiveness. It is difficult to be categorical about the effectiveness of these techniques, as the lack of open source data makes them difficult to use and rather unreliable.
Impact measurement enhanced by artificial intelligence sentiment analysis
The main objective of social economy organizations is to create social value, but this is difficult to quantify because of the complexity of the impacts produced by this sector. It is a complex and costly process, both economically and in terms of human resources, yet it is at the heart of the improvement and innovation of the social economy sector.
Impact analysis allows them to improve their practices and better communicate their impact to their stakeholders. This is an issue that is at the heart of the use and development of artificial intelligence. These tools already enable large-scale analysis of the feelings expressed by stakeholders. Using data collected on social networks for example, this would allow to identify trends on the impact of programs, services or fundraising efforts but also how they are perceived in the population.
The development of artificial intelligence specifically designed to analyze how a sample of the population feels about an action could be a solution to the difficulties faced by social economy actors on this topic.
AI development in the SE: the internal opinion of an expert
We had the chance to talk with a national expert in data science and artificial intelligence working for a large mutual. As a developer of artificial intelligence tools himself, he offered us an inside look at the current state of development of these technologies in the social economy. He clearly saw recent progress in a sector that he still defines as “difficult to move”. For him, the social economy enterprises have difficulty in moving forward in the adoption of data science and artificial intelligence, in particular because of the negative vision that these fields suffer from.
“For a long time the topics of artificial intelligence in the social economy have been under-considered and seen as only for ‘geeks’.”
The social economy also faces obstacles in terms of data collection. In fact, according to our expert, the main difficulties encountered by social economy actors are in the implementation of data collection processes and the delivery of results and predictions. In the mutual field, risk management is mainly based on business experience, so it is difficult to change methodologies to implement efficient and automated tools.
Despite these obstacles, artificial intelligence can bring many benefits to the social economy sector. The tools developed by the mutual insurance company for which our expert works help to optimize customer relations by better understanding the reasons for contact, sorting emails and scanned paper documents, and automating routine requests. These tools can also more easily detect fraud and thus improve the company’s efficiency and profitability.
“The results of these investments over the past decade are already palpable, with time savings and significant improvements in customer relations.”
In conclusion, the use of artificial intelligence in the social economy sector is a topic that raises important ethical questions, particularly in terms of data collection and use. However, artificial intelligence can offer opportunities to improve the efficiency and sustainability of fundraising campaigns, as well as to measure the impact of actions undertaken. While the literature in France on this topic is limited, companies and experts are already working on artificial intelligence solutions specifically designed for the social economy sector.
Convinced by this usefulness, Dr. Lobna Karoui, a Forbes board member and lawyer specialized in the ethics of artificial intelligence, gave an overview of the applications of artificial intelligence in the social economy sector. Whether it is to raise funds, improve the efficiency of their actions or better understand the needs of their beneficiaries, she believes that artificial intelligence is an asset for organizations. However, she warns that AI cannot and should not replace the human qualities of actors in the social economy.
The sector is evolving and there is no doubt that it will rise to the societal challenges raised by artificial intelligence. It is important to ensure that these tools are developed in an ethical and responsible manner, in order to preserve the integrity and values of social economy organizations.