Since the arrival of ChatGPT on the market in November 2022, almost everything has been said about the risks and dangers of these generative artificial intelligence (GAI) tools: massive job cuts, loss of creativity, false profiles , false images and information, etc. Much has also been said about the advantages, particularly economic, of this AGI for businesses. Fast in processing information or capable of translating into all languages, it frees up time which can be devoted to tasks with higher added value, such as automatically writing the minutes after a meeting.
But are these risks and benefits the same for women and men? Nothing is less sure ! The first inequality is that of numbers in the digital sector – and therefore AI –, where women are under-represented. Although relatively gender-neutral, professions linked to technology in general and digital technology in particular suffer from the under-representation of women in scientific and technical fields.
The “Pact for an egalitarian artificial intelligence between men and women”, launched by the Equality Laboratory, notes that women currently represent only 12% of jobs in the sector, and highlights that their absence “is one of the key reasons for the sexism of algorithms designed and developed by and in a male universe”.
The main risk of this under-representation is that of the gradual invisibility of women in many digital professions. At a time when IAG is rapidly becoming established at all levels of companies, “Women must be represented and active in all professions linked to AI, in particular to debias algorithms, include all diversity, remove stereotypes…insists Hélène Deckx van Ruys, CSR director and co-pilot of the AI group at the Equality Laboratory. Otherwise, biases will persist and be amplified”.
A complex situation
This is already the case in a little-known area of the world of work, that of the preparation of the data on which AGI models are fed. Paola Tubaro, research director at CNRS, specialist in the economics of digital platforms, was interested in shadow workers, who carry out microtasks on digital platforms to train tools like ChatGPT, everywhere in the world and particularly in Africa.
“These are gender-neutral tasks, which require just basic skills and which can be carried out by anyone, even at home, which could suit many women. However, there too, they are in the minority. Far from reducing inequalities between men and women, this work exacerbates them, she summarizes. Not only do they add an additional burden to women, who already manage domestic and family life, but because they do it occasionally, they perform lower-level tasks and earn less than men. » As a result, data preparation centers in French-speaking Africa today employ mainly men.
You have 50% of this article left to read. The rest is reserved for subscribers.