A close up shot of a small boy aged about six years old, with a dirty face. He has dark brown or black hair, big brown eyes and brown skin. His expression is sad or pensive. He is wearing a blue shirt. Much of the work that goes into creating generative AI tools is dangerous and/or harmful to the people involved. For example:

Content moderation work is carried out in the labour markets of countries such as Kenya or Uganda, where people are paid low wages to classify texts and images with violent, murderous, rape or child abuse content, and do not receive mental health support when encountering disturbing content on a daily basis
Fine tuning work on the algorithms underpinning generative AI tools is often outsourced, requiring people to manually score answers. The outsourced companies have been documented as operating in refugee camps, training displaced people in Lebanon, Uganda, Kenya and India, exploiting their economic hardship
The mining industry which supplies major digital hardware manufacturers involves labour from mines in places like Peru, Brazil and Congo, where there are issues like inequalities in the distribution of mining revenues, water pollution leading to health problems for local populations, illegal mines located on indigenous land, child labour and other human rights violations

Acknowledgement: Estampa (2024) 'Cartography of generative AI'. Available at: https://cartography-of-generative-ai.net/ (Accessed 1 May 2024).

Image: AshutoshGoyal on Pixabay