AI "Companion Bots" Actually Run by Exploited Kenyans, Worker Claims

Testimony to the Data Workers’ Inquiry from Kenyan worker Michael Geoffrey Asia says he was hired by subcontractor New Media Services as a “text chat operator” who impersonated chatbots in intimate conversations—often juggling three to five fabricated personas per day, meeting a 40 words‑per‑minute expectation, paid $0.05 per message and bound by NDAs. The account highlights that a widely used consumer niche (about 28% of Americans report intimate interactions with chatbots) is underpinned by low‑paid, high‑stress gig labor in developing countries, a supply‑chain reality (estimates of 154–435 million online gig workers) that creates reputational, regulatory and operational risks for AI firms and platforms reliant on opaque subcontracting models.

Analysis

Testimony from Kenyan worker Michael Geoffrey Asia to the Data Workers’ Inquiry alleges he was employed by subcontractor New Media Services as a “text chat operator” who impersonated chatbots in intimate conversations, juggling three to five fabricated personas per day, typing at least 40 words per minute, and earning $0.05 per message under a mandatory NDA. The account directly connects a widely used consumer niche — the article cites that 28 percent of Americans report intimate interactions with chatbots — to low-paid, high-stress gig labor in Kenya and similar markets.

Estimates in the article place the global online gig workforce between 154 and 435 million, and it highlights concentration of data labeling and moderation work in underdeveloped African, South American, and Southeast Asian nations. That supply-chain opacity creates clear reputational, operational, and regulatory exposure for platforms that outsource human-in-the-loop activities without transparent auditing or labor protections.

Sentiment and market-data signals show a moderately negative tone and a modest market impact score (0.28), implying near-term reputational risk rather than systemic market disruption. Potential investor-relevant outcomes include increased regulatory scrutiny, litigation or consumer-trust erosion for exposed consumer-facing AI firms, and possible margin pressure if companies internalize moderation or raise subcontractor pay.

Key Decisions for Investors

• Monitor regulatory inquiries, class actions, and company disclosures on human-in-the-loop practices and subcontractor audits for consumer-facing AI platforms
• Avoid initiating or increase exposure to AI/consumer platforms that do not disclose audited labeling or moderation supply chains and consider reducing positions where remediation timelines are unclear
• Favor companies that demonstrate transparent labor practices, onshore moderation, or technical investments in safe automation, and consider short-term hedges for names with concentrated subcontracting risk