The article discusses the use of AI in Amazon’s Mechanical Turk service, which was originally intended for human tasks. Researchers have found that nearly half of the workers on the platform are using AI, such as large language models like ChatGPT, to automate their work. This raises questions about the reliability of the platform and the integrity of the tasks being performed. The researchers caution that tasks like abstract summarization are particularly susceptible to automation. The article emphasizes the need for platforms, researchers, and crowd workers to find new ways to ensure that human data remain human in the face of advancing AI technology.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| Mechanical Turk workers using AI for tasks | Automation of tasks | Increased reliance on AI for completing tasks | Desire for increased productivity and income |
| Automation of Mechanical Turk tasks | Shift from human labor to AI labor | Decreased need for human workers on platforms like Mechanical Turk | Speed and reliability incentives |
| AI-generated data crisis | Crisis of AI training on AI-generated data | Need for new ways to ensure human data remains human | Advancement of AI technology and models |
| Threat of AI eating itself | AI self-referencing and misinformation | Increased skepticism towards automated content | Lack of certainty in distinguishing human work from AI work |