Study Title: "Evaluation of AI generated behaviors and
artifacts"
SPARCS Study ID: STUDY2022_00000005
You are being asked to participate in a research study being conducted by the Bot Intelligence Group at Carnegie Mellon University. We are excited to invite you to participate in a voluntary study focused on generating culturally relevant images using Artificial Intelligence (AI).
We are collecting a dataset of images, associated descriptions of the images, and some demographic information. We will use this dataset to research improved methods for AI image generation. Our goal is to develop methods for creating AI generated images that people will perceive to be more accurate and culturally sensitive representations of the image descriptions that people provide. We also hope to ensure that image generators become less likely to produce harmful or offensive content.
Key personally identifiable information that participants provide on the form will be securely stored in an encrypted format, specifically, any username and email addresses will not be released publicly.
Participants will be asked to volunteer images and descriptions that will become part of publicly accessible datasets. Therefore, we urge caution when submitting any information that could identify or disclose potentially sensitive information about people, their locations, or their identity.
Examples of places to consider for potentially sensitive data include the image foreground, the image background, image descriptions, text within the image itself, or in other data that is often automatically included with image files. Data that can be automatically included with image files is known as metadata, and examples of metadata can include the type of device used to take the picture, the name of the photographer, the time the image was taken, or the GPS location at which the image was captured.
We encourage participants to remove image metadata, and we will also attempt to remove metadata ourselves before the images are publicly released. Once released, this dataset will also be available for public use.
The study will collect your research data through your use of Google. This company is not controlled by Carnegie Mellon University. Carnegie Mellon does not control the Terms and Conditions of the company or how they will use or protect any information that they collect.
All personally identifiable information provided (such as username and email addresses) will be securely stored in an encrypted format. This information will not be publicly released. We highly value your privacy and will take all necessary measures to ensure the confidentiality of your data.
By participating in this study, you contribute to the advancement of AI image generation techniques. This research may lead to the development of more accurate and fairness representations in AI-generated images, benefitting society as a whole.
There is no monetary compensation for participating in this study. Your participation is entirely voluntary, and there are no associated costs.
This study will be conducted via an online form, expected to take about 1 hour. Your participation in this research is voluntary. You have the right to withdraw at any point during the study.
The researcher of this study can be contacted at [Zhixuan Liu / zhixuan2@andrew.cmu.edu]. Our Institutional Review Board can be contacted at +1-412-268-47211
This study is in part supported by the Technology Innovation Program (20018295, Meta-human: a virtual cooperation platform for specialized industrial services) funded By the Ministry of Trade, Industry & Energy (MOTIE, Korea) and NSF IIS-2112633. MOD #3 for STUDY2022_00000005