A prominent AI ethicist is urging a fundamental shift in how society perceives artificial intelligence, arguing that recognising its 'humanity' is essential to navigating its future. Eleanor Drage, an assistant research professor at the University of Cambridge’s Leverhulme Centre for the Future of Intelligence, contends that AI, far from being a mysterious, disembodied force, is intrinsically linked to extensive human labour – from the mining of raw materials to the meticulous process of data labelling.
Drage's perspective challenges the prevailing narratives propagated by Silicon Valley, which often oscillate between visions of AI as either a saviour ushering in an era of unprecedented prosperity or an existential threat capable of wiping out humanity. She argues that these extreme views, particularly the cultural obsession with AI-induced apocalypse, conveniently divert attention from more practical and immediate concerns. Instead, Drage advocates for critical conversations about making AI safer and more equitable, emphasising the need for citizens to gain greater control over how their personal information is used and how AI models are developed and regulated.
The academic's work frequently criticises big tech's practices, echoing concerns about algorithms perpetuating existing biases and the perceived lack of accountability from corporations that often make broad promises of transparency. Her research has previously debunked the efficacy of AI-powered recruitment tools and highlighted the risks associated with integrating AI into law enforcement. Drage posits that a handful of executives have amassed excessive power, driving AI deployment without adequate public consent or consideration for planetary health.
However, while Drage's philosophical explorations into the definition of intelligence are noted for their intellectual depth, some of her arguments have drawn scrutiny. For instance, her assertion that the labour force is not disappearing but merely shifting, from artists to data labellers, has been questioned for its lack of nuance regarding the differing socio-economic realities of these roles. Similarly, her dismissal of unconscious bias in the context of promoting diversity in AI has been described as potentially oversimplifying a complex issue.
Despite these criticisms, Drage's underlying goal is to present an alternative vision for AI development, moving away from top-down, profit-driven models towards more ethical, community-centred initiatives. She offers examples such as Mumkin, an app designed to facilitate difficult conversations around female genital mutilation in India, and Kuini, an AI chatbot supporting Māori women in their efforts to quit smoking. While these projects aim to address significant societal issues, questions have been raised about the specific value proposition AI brings to these endeavours.