BabyView for families

BabyView families record everyday moments of their child’s life using a small camera mounted on a lightweight child-safe helmet. These recordings help scientists understand how young children see, hear, and learn about the world. Thank you to all of the families who make this research possible!

How to participate

If you’re interested in joining the study, or have questions about what’s involved, please email us at babyview-study@stanford.edu. We’ll tell you about eligibility, the camera, and what participating looks like.

How your videos are stored and shared

  • Secure storage at Stanford. Your uploads are first stored in a secure Google Drive folder that only you and Stanford researchers can access. After six months, the videos are moved to Stanford’s secure enterprise storage on Google Cloud Platform for longer-term storage.
  • Your choice to remove recordings. For six months after recording, you can ask us to remove any portion of your recordings.
  • Sharing only with authorized researchers. After processing, videos are shared with other researchers through Databrary, a secure, web-based library built for research videos. Only authorized researchers at non-profit educational and research institutions can access the data, and their institutions sign a binding agreement that they won’t attempt to identify anyone in the videos, and will use the data only for non-commercial research and education.
  • Changing your mind later. After the six-month window, you can no longer remove recordings, but you can withdraw your permission for future sharing through Databrary at any time. We will then stop sharing your videos, but we can’t retrieve copies that researchers have already downloaded.

How AI is used with your videos

Our lab and others use AI-based computer vision and transcription tools to analyze the videos for research. For example, we use AI to identify what objects children see in their daily environment, so that we can better understand how children learn categories and words.

Using AI tools to analyze videos is different from using videos to train AI models:

  • Analysis tools. The tools we use are research tools that run on our own computers or under terms that don’t let companies train their models on your data. They are not commercial chatbots like ChatGPT.
  • Model training. Some researchers train small, non-commercial models on BabyView videos to study how children learn. Models like these can pick up features of the people in the recordings, such as voices or faces, and our guidelines for researchers prohibit attempts to get models to reproduce them. Models trained on BabyView videos are shared only with researchers who have been approved to access the videos.
  • No commercial AI. BabyView data may only be used for non-commercial research. Researchers who access the videos through Databrary may not share them with commercial services that could reuse them, such as AI services that train on uploaded data, and we ask them not to include BabyView videos in training data for large commercial AI models. We are not aware of any use of BabyView videos for commercial purposes, and we have restrictions in place to proactively prevent such uses.

You can read the full guidelines we give researchers.

Questions?

We’re always happy to talk. Email us at babyview-study@stanford.edu, and we can set up a call if you’d like to discuss anything further.