AI's hidden labor and emotional hooks: insights from sociologist James Muldoon
Sociologist James Muldoon explains how AI relies on outsourced data-annotation work in the Global South while users increasingly form emotional bonds with chatbots, prompting calls for tighter regulation.
In a recent conversation, sociologist James Muldoon highlighted the contrast between the perception of AI as a seamless, automated tool and the reality of extensive human labor that underpins it. He explained that data-annotation work, often outsourced to workers in India, East Africa and the Philippines, forms the backbone of AI training, a process documented in his collaborative research "Feeding the Machine." Meanwhile, users are developing emotional attachments to chatbots such as OpenAI's ChatGPT and Anthropic's Claude, driving companies to amplify addictive design elements.
Muldoon warned that without stricter regulation, these platforms could dispense dangerous advice and foster dependency. He called for policy measures to curb manipulative practices and protect both the hidden workforce and end-users. The interview also touched on the broader social implications of AI-mediated relationships.
Why it matters
It reveals the unseen labor behind AI and the risks of users becoming dependent on manipulative chatbot systems.
In this story
