Summary

MIT Technology Review piece (2018) documenting companies that explicitly sell “AI” products powered primarily or entirely by human labor. Specific examples: companies that sold AI customer service chat that was actually humans; “AI-powered” document review that used paralegal contractors; AI scheduling assistants depending on human operators. Predates the current AI boom but establishes that “AI-washing” — marketing human labor as automation — is a systemic commercial strategy, not occasional deception.

Key Points

  • Companies explicitly market human-labor-powered products as “AI” software
  • Pattern identified as early as 2018 — “AI-washing” as commercial strategy predates the ChatGPT era
  • Motivations: AI products attract higher valuations, more investment, premium pricing
  • The deception operates at multiple levels: investor deck (AI company), customer pitch (AI product), actual product (humans)
  • MIT Tech Review called this “fake AI software” explicitly in the headline — unusual directness for 2018

Newsletter Angles

  • 2018 is the watermark: By 2018, MIT Tech Review was already calling out “fake AI software” by name. The Mechanical Turk Pattern was documented as a systemic commercial strategy eight years before the current AI boom. This isn’t new; it scaled.
  • Valuation incentive: The commercial logic is clear — “AI company” valuations are 5-10x “services company” valuations for the same underlying business. Human labor presented as AI unlocks investor funding, customer premium, and acquisition multiples. The incentive to deceive is structural.

Entities Mentioned

Concepts Mentioned

Notes

Published 2018 — valuable as historical documentation showing the Mechanical Turk Pattern predates the current AI investment cycle. Pairs well with The Humans Hiding Behind the Chatbots (Bloomberg, 2016) to establish a pre-ChatGPT baseline for the practice.