Manage episode 515041134 series 3294843
Who teaches AI how to see and recognise? Definitely not a techie in a cool office in Silicon Valley. Most likely, it is a homemaker from a marginalised community in rural India who does this tedious work for very low pay. In his debut feature film "Humans in the Loop," director Aranya Sahay explores this vital detail that got left out of the "AI Revolution" marketing materials. The film follows Nehma, an Adivasi woman from Jharkhand who takes a job labelling data for AI systems. Turns out, teaching machines to recognise the world is a lot like raising a kid; except this kid absorbs prejudices from centuries of Western datasets, with no plan to augment it with context-specific knowledge. Esp. not from the labellers :-). The film has won major accolades, including Best Film at the Bengaluru International Film Festival and Grand Jury Prize for Best Feature at the Indian Film Festival of Los Angeles. Have fun watching the episode! Nishtha Manchanda: https://www.linkedin.com/in/nishtha-manchanda-3b6b421b? 00:00 Introduction to the Film's Context 01:40 The Paradox of AI and Simple Tasks 02:21 Government and Civil Society's Role in AI 02:55 Future of Technology and AI 03:42 About the film 'Humans in the Loop' 04:32 Exploring Data Labeling and Its Implications 06:12 AI as a Child: A Philosophical Perspective 08:05 Challenges and Realities of Data Labeling 13:00 Bias in AI and Its Consequences 17:51 AI in the Film Industry 23:36 Global Reactions to AI and the Film 26:04 Ethnographic Insights and AI 26:27 Tech Reactions to the Film 27:46 Unexpected Audience Reactions 28:09 Women and the Emotional Impact of the Film 29:57 Community and Cultural Representation in AI 32:45 Government and Civil Society's Role in AI 35:41 Future Projects and Superhero Films 36:05 The Journey of the Film 36:50 Story Culture and Film Funding 39:08 Indigenous Life and Environmentalism in Film 40:46 Superhero Films in India 42:18 Film Screenings and Audience Engagement 43:50 Concluding Remarks and Future Endeavors
Learn more about your ad choices. Visit megaphone.fm/adchoices
88 episodes