Richard Mathenge

Richard Mathenge is the Kenyan tech-worker organizer and former Samasource team lead who, among the four public petitioners on the July 2023 parliamentary petition against OpenAI and Sama, has developed the most distinctly policy-advocacy and academic public-output register. Tracked here as a Voice because his output since 2023 — podcast testimony, a Stochastic Parrots Day panel appearance, co-authored policy writing in The Conversation, academic co-authorship, and a seat on the Partnership on AI Policy Steering Committee — is the load-bearing object the corpus needs to track beyond the affiliation record in his Person entry. Biographical and affiliation detail, including his co-founding of the African Content Moderators Union (ACMU) and Techworker Community Africa (TCA), are recorded there.

Voice anchor

Mathenge's Voice closes the "organizer-as-policy-advocate" slot in the corpus's Kenya AI labor cluster — a register distinct from the personal-testimony voice that Mophat Okinyi anchors.

• Distinct from Okinyi's testimonial register. The corpus's Kenya AI supply-chain cluster already carries Mophat Okinyi's Voice entry, which anchors the "invisible workers" personal-disclosure register — the most extensively reported individual account of the psychological toll of the work. Mathenge's register is the organizer-turned-policy-advocate counterpart: where Okinyi's public-output practice runs primarily through personal testimony and rights education, Mathenge's runs through formal co-authored policy and academic writing, institutional engagement, and regulatory advocacy. The two registers complement rather than duplicate each other within the same organizing cluster.

• First AI supply-chain worker voice in the corpus at a major multistakeholder AI table. Mathenge's October 2023 appointment to the inaugural Partnership on AI Policy Steering Committee — under his ACMU affiliation, alongside Rumman Chowdhury, Alexandra Givens, Sam Gregory, Alondra Nelson, and Irene Solaiman — is structurally unique in the corpus's voices slice: he is the only AI supply-chain worker voice who holds an institutional seat at a major AI multistakeholder body while simultaneously building worker-organizing vehicles and co-authoring formal policy and academic writing on the labor question. This floor-to-table trajectory — from Sama team lead at the RLHF production floor to PAI Policy Steering Committee — is a deliberate organizing posture, not an incidental credential.

• First academic co-authorship from the worker-organizer layer. The December 2025 arXiv paper "Mapping Data Labour Supply Chain in Africa in an Era of Digital Apartheid: a Struggle for Recognition," co-authored with Pidoux, Kypraiou, Kgomo, Malgwi, Okinyi, Oyange, and Tighanimine, is the first instance in the corpus of an AI supply-chain worker-organizer appearing as a named academic co-author on a formal mapping study. Its use of Honneth's "struggle for recognition" framework to argue that content moderators deserve recognition of their expertise rather than treatment as unskilled workers is the theoretical articulation of what Mathenge has argued on the organizing and testimony register since 2023.

Signature framings

Three arguments anchor Mathenge's public-output register.

• Regulatory intervention as the structural fix. In the March 2025 Conversation article co-authored with Mohammad Amir Anwar, Mophat Okinyi, Kauna Malgwi, and Sonia Kgomo, Mathenge advances a four-part regulatory framework: continent-wide BPO accountability frameworks with penalties for labor violations; job-quality-over-quantity policies that push African governments toward structural diversification rather than sector expansion; robust social protection mechanisms; and naming-and-shaming transparency tactics combined with collective bargaining. This is the policy-advocacy register in its clearest form — naming specific institutional interventions, not just documenting harm — and distinguishes Mathenge's voice from the primary-testimony register that dominates the corpus's other Kenya AI labor entries.

• "For too long we, the workers powering the AI revolution, were treated as different and less than." This formulation, delivered at the DAIR Stochastic Parrots Day panel "On Worker Exploitation, Data Theft, and the Centralization of Power" (May 2023) alongside Safiya Noble, Milagros Miceli, and Timnit Gebru, is Mathenge's canonical universalizing claim. The framing connects the Kenyan content-moderation labor case to a global argument about how AI development systematically devalues the workers whose hand-classifications make frontier-AI safety systems possible — not as an individual grievance but as a structural wrong. Appearing in the same month as the ACMU founding vote, it marks the organizing moment's transition from testimony to political claim.

• Data transparency as a labor rights tool. Mathenge's 2025 personal data access request to his former employer Teleperformance — documented in Rest of World — advances a distinct regulatory tactic: using GDPR-style data access rights to surface what labor documentation the outsourcer holds about workers. The inadequate response he received — very little information, some inaccessible — is itself evidence that worker data is being withheld, turning the access request into a documentation act. This registers a new accountability axis alongside the parliamentary petition and union organizing channels: using data-protection law as a labor-rights tool.

Public output and venues

Mathenge's public-output register runs through four overlapping channels.

• Big Technology Podcast and early US media testimony. His Big Technology Podcast appearance with Alex Kantrowitz, "He Helped Train ChatGPT. It Was Traumatizing." (May 2023, 47 minutes), was one of the first sustained English-language media accounts of the RLHF labor conditions from a named Kenyan team lead — placing the specific shift structure (nine hours per day, five days a week of explicit-content classification), the pay rate (under USD 2 per hour), and the psychological consequences into the US tech-media register before the June 2023 TIME100 AI nomination and the July 2023 parliamentary petition. The podcast appearance established his public testimony register in the format most accessible to US-based AI policy audiences.

• Academic AI ethics venues. His Stochastic Parrots Day panel at the DAIR Institute event (May 2023) — "On Worker Exploitation, Data Theft, and the Centralization of Power" — placed Mathenge alongside Safiya Noble, Milagros Miceli, and Timnit Gebru in the academic AI ethics venue that the Stochastic Parrots paper's anniversary convenes. This register connects the Kenyan worker testimony to the scholarly critique of AI's structural labor exploitation, positioning Mathenge as a practitioner voice in an academic venue rather than exclusively as a media source.

• Co-authored policy and academic writing. The March 2025 Conversation article and the December 2025 arXiv paper mark Mathenge's transition from media-sourced testimony to named-author policy and academic writing. The Conversation piece is op-ed format directed at policy audiences; the arXiv paper is a social-science mapping study covering Africa's content moderation industry across 43 countries. Both are first-person-as-expert rather than first-person-as-witness — the organizer-as-authority shift that distinguishes his register from the testimony layer.

• Documentary, institutional engagement, and oral history work. His data-workers.org documentary project — examining working conditions of content moderators and data annotators at Sama in Nairobi — aims to expose behind-the-scenes conditions and inspire workers globally to collectively advocate for their rights. His Partnership on AI Policy Steering Committee seat and interviews including the Creative Change Makers conversation (January 2024) round out a practice that moves between the organizing ground and the institutional policy table.

Why this is a Voice entry

A Voice entry is created here because Mathenge's public output — podcast testimony, academic AI ethics panel appearances, co-authored policy and academic writing, institutional committee work, and documentary production — is the load-bearing object the corpus needs to track beyond his organizational affiliations. The Person entry records his ACMU administrator role, TCA co-founding, parliamentary petition participation, and Partnership on AI committee seat; this Voice entry tracks the public-output practice through which he has advanced the argument that AI supply-chain workers deserve regulatory protection and recognition as skilled workers, not just documentation as victims. The corpus's Kenya AI labor cluster had Mophat Okinyi's personal-testimony register; this entry adds the organizer-as-policy-advocate register and the first academic co-authorship from the worker-organizer layer.