Covered But Not Protected: The Risks Behind Online Sharing

by Jennifer Pruitt

September 18, 2025 / 12:45 PM CST

In the age of social media, it has become second nature to share milestones and memories online. Birthdays, game days, and exciting moments often find their way onto social media within minutes. We all want to share our lives, celebrate wins, and mourn losses in solidarity with friends and family, especially when loved ones live far away. To balance that desire with internet safety, many parents blur parts of an image or place an emoji or a sticker over their child’s face. While this may seem effective, this practice is not as safe as it seems.

Artificial Intelligence (AI) and digital tools are quickly advancing and are now accessible to everyday users. These technologies can be used to sidestep overlays or generate an image from what remains in the photo. Even without these technologies, every photo posted contains small details. Individually, these details seem harmless, but when pieced together across multiple posts, those details can reveal more about your child than you ever meant to share.

Blurring, pixelating, or low-resolution masking has been the media’s go-to privacy method for years. Pixelating an image isn’t new, or even complex, technology and it has long provided a sense of security because the human eye cannot see through it. For a while, that was all that was needed. Today, however, modern technology has the capability to see through this privacy screen and unveil what is underneath.

Researchers at the University of Texas at Austin and Cornell Tech demonstrated that software could be trained to identify objects that have been pixelated or blurred. According to Cornell Tech’s Vitaly Shmatikov, the training methods used were very standard in image recognition techniques, widely known and easily accessible. By processing massive datasets, the software learned how features typically appear, even under heavy pixelation or blur. Shmatikov warns that bad actors with minimal technical knowledge could carry out this type of attack.

Many parents have moved from blurring or pixelating images to placing an emoji or sticker over portions of the photo. While these overlays seem like a safer option, it often creates a false sense of security due to other details left exposed in the photo.

Although there are no verified cases of emojis or stickers being successfully removed from a photo, experts still caution that information can be gathered from the photo even without showing facial features. Lisa Ventura, the founder of Cyber Security Unity, said “putting an emoji over a child’s face provides virtually no real privacy protection whatsoever . . . most parents aren’t just posting one carefully emoji-protected photo,” instead they are posting “multiple images over time.”

Viewed collectively, these images can reveal information about the child that parents don’t intend to reveal. Photos, even with an overlay, can reveal things like birthday, age, body build, location, personal interests, or even the school they attend. This information can be gathered not only from the photo, but also from captions, photo metadata*, backgrounds, what your child is wearing, and even shadows. Advanced tools can be used to analyze patterns from multiple photos to generate a face that is very similar to a child. Even without a clear picture, a bad actor could identify, locate, and approach a child using the personal details gleaned from those posts.

While it seems like blurring, pixelating, or placing an emoji or sticker over an image would safeguard privacy, that just isn’t true in today’s digital world. What once passed as a quick fix for internet safety simply doesn’t work against modern tools.

So, what can we do to safely share moments with each other? You don’t have to stop sharing moments, instead, make small adjustments to protect you and your child’s privacy:

-          Change the angle: Take photos from behind, above, or at a distance so that your child’s face isn’t visible.

-          Details, not faces: Share photos of your child holding a school project, their feet on the baseball field, or the birthday decorations instead of your child’s face.

-          Check the background: Be mindful of house numbers, landmarks, signs, or logos that give away location.

-          Tighten privacy settings: Change your setting so that photos can’t be downloaded. Share only with close family and friends through direct message, closed group messages, or private accounts.

-          Share offline if possible: Text or email photos of special moments to loved ones instead of posting them publicly.

-          Crop photos before posting: Remove details entirely instead of using overlays and consider sharing screenshots to strip out metadata.

These simple changes allow you to continue celebrating life’s moments while keeping your family safer. You don’t have to stop sharing, just share smarter.

Want to learn more or schedule a training on this topic? Reach out to Jennifer Pruitt, Outreach Specialist, at jpruitt@cacpalopinto.org.

*Metadata: the background information that describes a file, photo, document, or message. This may include the date and time the file, photo, or document was created, GPS location, the creating device type, editing history, author’s name, etc.

References:

1. Jonathan Bright, Florence E. Enock, Saba Esnaashari, John Francis, Youmna Hashem & Deborah Morgan, “Generative AI Is Already Widespread in the Public Sector,” arXiv:2401.01291, January 2, 2024, https://arxiv.org/abs/2401.01291.

2. Longtao Jiang, Zhendong Wang, Jianmin Bao, Wengang Zhou, Dongdong Chen, Lei Shi & Houqiang Li, “SmartEraser: Remove Anything from Images using Masked-Region Guidance,” arXiv preprint arXiv:2501.08279v1, January 2025, https://arxiv.org/html/2501.08279v1.

3. Richard McPherson, Reza Shokri & Vitaly Shmatikov, “Defeating Image Obfuscation with Deep Learning,” arXiv:1609.00408 [cs.CR], September 2016, https://arxiv.org/abs/1609.00408.

4. Lily Hay Newman, “AI Can Recognize Your Face Even If You’re Pixelated,” WIRED, September 12, 2016, https://www.wired.com/2016/09/machine-learning-can-identify-pixelated-faces-researchers-show/

5. Chas Newkey-Burden, “Sharenting: does covering children’s faces on social media protect them?” The Week UK, July 2, 2025, https://theweek.com/culture-life/personal-technology/sharenting-covering-childrens-faces-on-social-media-emojis.

6. Julian Todt, Simon Hanisch & Thorsten Strufe, “Fantômas: Understanding Face Anonymization Reversibility,” Proceedings on Privacy-Enhancing Technologies 2024, https://petsymposium.org/popets/2024/popets-2024-0105.pdf