AI-Shaped Inequalities

The chapter examines AI inequalities through the lens of the digital divide debate, emphasising their relation to pre-existing social and digital inequalities. Although most AI tools are free, an AI usage gap has already emerged, whereby children from disadvantaged backgrounds use AI tools less for learning and school. Moreover, vulnerable children are more likely to turn to AI companionship chatbots, with potentially adverse outcomes. AI inequalities pertain also to the second-level digital divide, namely literacy. In fact, preliminary data show how AI knowledge and skills are unevenly distributed across age, education, digital skills and socioeconomic status. Inequalities in access, usage and literacy translate into an unequal distribution of AI opportunities and risks. While the tangible outcomes of predictive AI encompass allocative and representational harms, and differential exposure to manipulative personalised advertising, differential Gen AI access, usage and literacy results in unequal educational, cognitive and wellbeing outcomes. Finally, the chapter explores the exploitation of children’s manual and cognitive labour in Global South countries that lies behind the illusionary autonomy, intelligence and artificiality of AI systems.