OpenAI is inviting independent contractors to provide actual work assignments from their job history. This effort is aimed at utilizing the information to evaluate the capabilities of its upcoming AI models, as revealed by documents obtained from both OpenAI and Handshake AI, a data training firm cited by WIRED.
This initiative appears to be part of OpenAI’s larger approach to create a human benchmark for various tasks, allowing for comparisons with AI systems. In September, OpenAI launched a new assessment method to evaluate how its AI operations measure up against human professionals in different fields. The organization emphasizes that this measure is crucial for tracking progress towards achieving Artificial General Intelligence (AGI), where AI may outperform human skills in economically beneficial tasks.
Contractors are encouraged to provide details on their work tasks and to upload substantial examples of their deliverables, such as Word documents or PDFs. An OpenAI presentation highlights that these exemplars should represent “real, on-the-job work,” with illustrations drawn from tasks like preparing an itinerary for a luxury yacht trip. The company stresses the importance of excluding sensitive and personal information from the documents shared, advising careful handling to prevent unintended exposure of confidential data.
In summary, this move reflects a trend within AI organizations to enhance their models’ effectiveness by leveraging the expertise of real-world tasks. Major players like OpenAI, Anthropic, and Google are increasingly turning to contractors for high-quality training data, thereby cultivating a profitable segment within the AI training sector. As the demand for quality intensifies, these firms may invest more resources in skilled individuals capable of meeting these standards.
The ainewsarticles.com article you just read is a brief synopsis; the original article can be found here: Read the Full Article…
