Hajra Naz
OpenAI and training data company Handshake AI are reportedly asking third-party contractors to upload real work they have done in past and current jobs, according to Wired. This is part of a growing trend in the AI industry.
Companies are hiring contractors to produce high-quality training data. The goal is to train AI models to handle more complex white-collar tasks. These tasks include things like drafting documents and analyzing data.
In OpenAI’s case, a company presentation reportedly instructs contractors to describe tasks they’ve performed at other jobs and upload examples of “real, on-the-job work” that they have personally completed. The requested materials can include concrete outputs such as Word documents, PDFs, PowerPoint presentations, Excel sheets, images, or code repositories—rather than summaries or descriptions of the work.
Contractors are reportedly told to remove proprietary information and personally identifiable data before uploading. OpenAI provides a tool called ChatGPT “Superstar Scrubbing” to help with this process, aiming to reduce the risk of sharing sensitive material.
Despite these safeguards, legal experts warn that the approach carries significant intellectual property and liability risks. Intellectual property lawyer Evan Brown told Wired that any AI company asking for real work in this manner is “putting itself at great risk,” because it requires trusting contractors to determine what is confidential or proprietary.
Analysts say this method shows a tension in AI development. High-quality training data is crucial for building models that can handle complex tasks. But using real work from contractors raises questions about consent, ownership, and trade secrets. Companies like OpenAI and Handshake AI are trying to balance these risks. They need data that is authentic and detailed enough to train models for real-world professional tasks.
Read More: The AI Leadership Race 2025-2027: Which Company Will Dominate?
Privacy advocates warn that even with scrubbing tools, sensitive information could slip through. This is especially true for legal documents, financial records, or code with embedded credentials. The practice highlights broader ethical and legal challenges for AI labs as they expand data collection from human contributors.
The move also shows a growing trend in the AI sector. Contractors, not full-time employees, are becoming the main source of high-quality training data. This raises questions about labor rights, fair compensation, and transparency in AI development.
