A claim that outsourced personnel, rather than in-house staff, reviewed users' ChatGPT conversations โ fueling a privacy-invasion controversy around OpenAI under the banner of "Project Lilly" (Korean: ํ๋ก์ ํธ ๋ฆด๋ฆฌ) โ has been put to a nine-source review. Direct capture of the original official documentation proved impossible after access was blocked, but the substance of the claim was judged to hold up on the whole.
The assertion circulating in Korean reads: "Outsourced workers review ChatGPT conversations," tied to allegations that OpenAI's "Project Lilly" infringed on user privacy. It rests on two points: that conversation review is carried out by outsourced personnel, and that the practice exposes users' chats to privacy risks under the project's name.
The review drew on nine primary sources. Its central limitation: direct, first-hand capture of the original official document was blocked, leaving the document itself unconfirmed. The ruling therefore rests on the weight of the remaining sources rather than on the document alone. The blocked access also means the exact spelling and finer particulars of the "Lilly" project name โ a transliteration from Korean โ could not be pinned down with certainty.
The core assertions tracked with the assembled sourcing: outsourced personnel are engaged to review ChatGPT conversations, and the practice has stirred a privacy controversy. Because the primary document could not be captured directly, however, the details tied to the project designation fall short of first-hand confirmation โ the gap that keeps the claim from a full endorsement.
Weighing the nine sources against the blocked document, the claim that outsourced workers reviewed ChatGPT conversations amid a "Project Lilly" privacy controversy is rated Mostly True.