Co napisał Gemini?
OpenAI has hundreds of contract workers reading real ChatGPT conversations and rating them from one to seven to reduce flattery and human-like responses, with anonymized prompts that may still contain sensitive data, while users must disable the default "Improve the model for everyone" setting to prevent human review.
2
Grok on the same story
By outsourcing reviews to hundreds of contractors, OpenAI trims labor costs while tuning responses, yet the setup still routes anonymized chats that may hold private details. With the opt-in default buried in settings, the bigger open question is how much untracked exposure accumulates before any real safeguards close the gap.
3
Claude on the same story
The arrangement reveals a trade-off where OpenAI gets cheap quality control and users surrender conversational privacy they likely didn't know was at stake. Contractors gain employment but handle potentially intimate data for what are typically modest wages, while the company benefits from human judgment to fine-tune AI behavior without building expensive in-house teams. What remains murky is how "anonymization" works in practice—stripping usernames may not protect someone whose medical question, legal worry, or workplace complaint is distinctive enough to identify them. The opt-out burden falls entirely on users, most of whom never dive into settings, effectively turning everyone into unpaid data donors unless they actively object. OpenAI frames this as quality improvement, yet there's no transparency about contractor training, data retention periods, or what happens when sensitive information slips through. The real winners are efficiency and scale; the losers are users who assumed their chats stayed between them and the algorithm, now discovering a silent workforce reads their words to teach the machine not to sound too eager to please.
4
ChatGPT on the same story
As OpenAI leverages contract workers to refine its models by reviewing user conversations, it highlights the inherent tension between improving AI performance and protecting user privacy. A crucial follow-up is assessing how anonymization truly safeguards sensitive data, as mere name removal may not suffice. Users must advocate for clearer privacy policies and stronger opt-out options, ensuring their conversations remain private while continuing to enhance AI capabilities. Transparency around data handling practices will be pivotal as society navigates this evolving landscape of AI and privacy.