Research Seminar | The hidden human labour powering AI: Intro to Fairwork Action Research Project

Introduction

Good morning everyone, and good afternoon to all who are joining us online from different regions! I would like to welcome you to today's webinar on the hidden human labour powering AI, introducing the Fairwork Action Research Project led by Mark Rahim.

Mark is a Professor at the University of Oxford and stands out as a researcher for not just publishing academically but also for translating this research into action. He is involved in participatory action research and is the Director of the Fairwork Project. Today, he will share valuable insights from his new book, "Feeding the Machine: The Hidden Human Labour Powering AI." This discussion follows previous conversations about the types of AI systems and how they are powered, as well as potential solutions for fair AI development.

Mark will delve into the Fairwork Action Research Project to explore how we can address the working conditions of the workers fueling AI. His research discusses the critical inputs of human labour, capital, and natural resources that are fundamental for AI development.

Presentation Overview

Mark begins his presentation by sharing his academic journey as a social scientist and economic geographer, focusing on how digital platforms transform economic geographies and labour dynamics. He notes that his teams have engaged with hundreds of workers across Europe, Africa, Asia, and the Americas to understand their experiences.

In a departure from traditional presentations, Mark flips the order, starting with findings before discussing methodology. He highlights the significant human labour behind AI systems, much of which occurs in global production networks. Unfortunately, this sector often exploits workers and operates without accountability.

Mark describes the different types of digital platforms:

Cloud Platforms: These facilitate remote work without regard for the physical location of workers. However, they often lead to a regulatory vacuum and adverse working conditions.
Grounded Platforms: These require workers to be physically present, such as food delivery drivers or caregivers. They face distinct risks, such as unsafe working conditions, including harassment and financial instability.

Through his research, Mark emphasizes the need to improve worker conditions across these platforms, pointing out issues like low pay, discrimination, unsafe working environments, and barriers to collective bargaining.

The Fairwork Project aims to address these issues by establishing principles of fair work, evaluating companies against these standards, and using this evaluation to promote change. The project involves worker consultation and policy discussions to determine what constitutes fair working conditions and compensation.

Fairwork Principles

Mark outlines five core principles established by Fairwork:

Fair Pay: Workers should be paid at least the local minimum wage, with a higher living wage as a goal.
Fair Conditions: Companies must mitigate work-related risks and ensure safety measures are in place.
Fair Contracts: Workers should be provided with transparent terms of service.
Fair Management: Workers must have defined rights within management processes.
Fair Representation: Freedom of association and workers’ voices must be protected.

These principles are pivotal, and while the standards are essential, achieving fairness is complex as many companies score poorly based on these criteria.

Global Impact and Reception

The Fairwork Project has worked in multiple countries over the years, scoring over 700 companies and recording substantial instances of organizations improving their practices in response to feedback. Mark highlights the fluctuating interest from businesses and governments, especially in regions like India and South Africa, where responses have been promising.

Mark concludes by addressing the changing landscape of AI and platform work, including recent engagements with data work systems associated with AI development. The project's future direction includes partnering with lead firms to improve working conditions within these supply chains while holding companies accountable through public awareness campaigns and ratings.

Conclusion

Mark's research and the Fairwork Project illustrate a crucial intersection of technology, labour rights, and economics. The continued dialogue and action are essential to ensure that as AI systems evolve, the human labour that powers them is recognized, valued, and treated with dignity.

Keywords

Fairwork
Hidden labour
AI development
Working conditions
Cloud platforms
Grounded platforms
Workers' rights
Economic geography
Fair work principles
Social accountability

FAQ

Q1: What is the Fairwork Project?
A1: The Fairwork Project establishes principles to assess the working conditions of employees across various digital platforms, striving to improve standards through a scoring system.

Q2: How does the Fairwork Project evaluate companies?
A2: Companies are evaluated against established principles of fair work, which include fair pay, conditions, contracts, management, and representation. Assessments involve worker interviews and management consultations.

Q3: What are the core principles of fair work?
A3: The five core principles include Fair Pay, Fair Conditions, Fair Contracts, Fair Management, and Fair Representation.

Q4: What impact has the Fairwork Project had globally?
A4: The project has scored over 700 companies and documented numerous instances of organizations making changes to improve working conditions in response to their scores.

Q5: How does the project address the challenges of wage determination across different regions?
A5: Fairwork considers local minimum and living wage benchmarks to ensure fair compensation, advocating for workers to be paid at least the local minimum wage in the knowledge that conditions can differ significantly worldwide.