In light of the growing use of generative AI and large language models (LLMs), PDRC has added a new provision concerning the use of generative AI and related technologies to its data use pledge, in order to further strengthen the appropriate management of data provided by PDRC and the protection of survey respondents’ information.

Under the new provision, users are prohibited from inputting, transmitting, or uploading unit-level data provided by PDRC, or any individual-, household-, or other survey-unit-level data created by processing or extracting such data, to any service or software equipped with generative AI, large language models (LLMs), or other similar functionality, or from allowing such service or software to process such data.

For the time being, the same restriction also applies to services provided under contracts with universities or research institutions, services that do not use submitted content for AI training, and institution-hosted or on-device LLMs that do not communicate externally.

On the other hand, generative AI may be used to create or modify analytical programs, provided that no unit-level data or other non-public information are supplied to the AI and only publicly available variable information is used.

The Data Use Pledge and FAQ, including this provision, are available on the PDRC website. Please refer to them for details on the provision and its specific application.

Users who are currently using data provided by PDRC will receive a separate email with instructions on the additional pledge procedure. Please review the newly added provision and indicate your agreement to comply with it.

Please note that users who do not agree to comply with this provision will not be able to continue using data provided by PDRC.

Thank you for your understanding and cooperation.