A young lad in England working with a thread-making machine back in 1909.
Amazon’s grocery stores made a significant change by discontinuing the “Just Walk Out” technology recently. Surprisingly, it was discovered that this automated checkout system involved the participation of 1,000 reviewers based in India. This instance is just one among many where advancements in automation were actually driven by concealed human intervention, a phenomenon famously referred to as “The Mechanical Turk.”
The Mechanical Turk originally denoted a deceptive chess-playing contraption dating back to 1770. This device seemingly operated autonomously, engaging in competitive chess matches with human opponents. For years, it captivated audiences worldwide as an early automaton marvel. However, the truth eventually emerged that a skilled chess master was clandestinely housed within the contraption, manipulating its moves.
This concept resurfaced in early 1800s England in a different context. As detailed by Brian Merchant in his book “Blood in the Machine,” textile manufacturers in England introduced new “automated” machinery to replace manual labor. Interestingly, these machines, though not reliant on skilled artisans, still necessitated the involvement of young children (often orphans) to operate them. Thus, these systems, while not fully automated, produced inferior quality products inexpensively, displacing skilled textile workers who were derogatorily labeled as “Luddites.”
Fast forward to the present era dominated by artificial intelligence, where remarkable feats of automation coexist with numerous instances of deceptive practices. Tech companies are in a race to showcase cutting-edge AI technologies, often prematurely implementing them. Consequently, many firms are shoring up pseudo-AI models with human labor.
The following are ten contemporary examples of Mechanical Turks, where automation is essentially a facade for human intervention:
-
Amazon’s Just Walk Out technology at grocery stores was recently phased out, revealing a reliance on a team of over 1,000 associates in India for its operation.
-
Presto Voice, a company offering drive-thru automation services to major brands, was found to require human intervention for over 70% of its orders, utilizing offsite workers in the Philippines.
-
Facebook’s virtual assistant, M, touted for its capabilities in 2015, primarily relied on human operators rather than advanced AI systems for tasks like booking movie tickets or ordering food.
-
Gemini’s AI demonstration in December 2023, showcasing real-time analysis of various inputs, was later exposed for its edited video and human assistance in generating responses.
-
Amazon Mechanical Turk, an online marketplace, enables developers to hire individuals for tasks challenging for computers, often paying minimal amounts for tasks that masquerade as automated.
-
Expensify’s expense reporting feature, which claimed automation, was revealed to involve human workers, including Amazon Mechanical Turk participants, in processing expenses.
-
X.ai’s virtual assistants, advertised as AI-driven, were actually operated by humans responding to emails, scheduling meetings, and performing various tasks.
-
Google’s practice of scanning Gmail inboxes for personalized ads, though discontinued, was replaced by human developers outsourced to review user emails for feature development.
-
Theranos’ automated blood testing device, a fraudulent venture by Elizabeth Holmes, was exposed for its lack of automation, with employees conducting tests using external lab equipment.
-
Scale AI, a prominent startup, relies on 10,000 remote workers in the Philippines to train its AI models, underscoring the substantial human involvement behind seemingly automated systems.
