Can Gaggle demonstrate the efficacy of its products?
Each year, we share our findings from the most recent school year in our annual Through the Gaggle Lens: The State of Student Safety report. This report dives into what we saw during the school year, the alerts we sent to districts about students in crisis, and the types of troubling incidents we flagged, offering insight into the state of student safety as well as the successful interventions resulting from our services. But don’t just take our word for it—read our Case Studies to learn more about how we partner with K-12 school districts to help enhance student safety and the results our district partners are seeing.
How does Gaggle generate and send alerts, and who receives the alerts?
Gaggle leverages a combination of artificial intelligence, machine learning, and a highly trained content review team to flag concerning content that K-12 students may receive, send, or create using school-provided technology. Gaggle uses a combination of keywords, algorithms, and machine learning to identify content that indicates students planning self-harm, bullying, abuse, or school violence.
If a concerning item is identified by Gaggle’s technology, a member of our trained human review team (the Gaggle Safety Team) will analyze the item to assess context and determine whether it is a false positive or if it might be something of concern. If it is determined that the content in question indicates a mental health issue, a threat, or potential harm to a student, it is escalated to a second Gaggle Safety Team member who will look more carefully at the item and full context, categorize the concern, and determine the severity and urgency.
From this point, the following actions occur:
- Non-Urgent Email Alert: If the second Gaggle Safety Team member determines the item is concerning but not urgent, an email alert will be sent to the emergency contacts at the school district.
- Urgent Telephone Contact: If the flagged item suggests an immediate threat to a student or someone else, the Gaggle Safety Team member will contact by phone the school district’s designated emergency contacts, who could be a superintendent, a counselor, a principal, or other school official. With guidance from Gaggle, schools set their own emergency contact processes and policies.
- Police Wellness Check (only in extremely urgent circumstances): If there is no response from the district-appointed emergency contacts after three contact attempts and a child’s life appears to be in imminent danger, Gaggle will reach out to the local authorities for a wellness check. Gaggle avoids contacting law enforcement directly unless there is an extreme circumstance where the Gaggle Safety Team is unable to reach the school-designated emergency contact and the flagged content indicates a child may be in immediate danger.
- Federal Law Enforcement Alert (only in instances of child sexual abuse materials): If the item contains an image that appears to portray a sexually explicit picture of a child, the item is registered with the NCMEC, as required by federal law. The district-appointed emergency contacts are notified about the existence of the image, but not sent the image in question.
What content does Gaggle review?
Gaggle provides a monitoring solution for Google Workspace for Education, Microsoft 365, and the Canvas learning management system. Items monitored include students’ school-provided email accounts, document creation and collaboration on school-provided educational platforms, calendar entries, chat, internet web activity occurring on Safari, Edge, and Chrome browsers, and other direct and group communication tools used as part of a learning or educational setting. Gaggle can also monitor activity occurring on local devices outside of the browser on Chrome and Windows computers.
Who can purchase the service?
Gaggle is not available for consumer purchase. Gaggle is only available for purchase to schools and districts to be used on school-issued devices and accounts.
How long does Gaggle retain data?
Gaggle purges non-incident data (i.e., data that does not indicate a threat of harm) after 30 days. Incident data is retained for individual school districts until one of the following actions occurs: (a) the relationship with Gaggle is terminated, (b) the student in question graduates or withdraws from the district, or (c) the school district requests a full data purge. All student information is anonymized.
Are Gaggle’s keywords and content categories for alerts appropriate and useful for the purpose of self-harm monitoring?
If our technology identifies something concerning—such as the use of a keyword, flagged phrase, or troubling image—that content is then reviewed by a Gaggle Safety Team member in order to verify the content, assess the context, and determine the level of concern. If our Gaggle Safety Team determines the flagged content is life-threatening, a member of the team immediately alerts the designated school emergency contact by phone.
Equitable prevention practices are paramount to providing students with a fair education and the support they need. Our algorithm reviews content anonymously, so we have no context or background on students when we identify potential issues. During onboarding and training for our Gaggle Safety Team, we include topics related to bias and opinion, as well as the importance of separating these from decision-making related to items they review.
Occasionally, districts identify trends or slang that may be unique to their community of students, and Gaggle will create additional monitoring filters to capture this content. We believe it’s important that the technology remain nimble and adapt to students’ ever-changing ways of communicating.
How does Gaggle recognize context?
If a concerning item is identified by Gaggle’s technology, a member of our trained human review team (the Gaggle Safety Team) will analyze the item. The Gaggle Safety Team undergoes extensive internal training with our experienced reviewers. The Gaggle Safety Team supervisors have an average of five years’ history with Gaggle analyzing content and additional training from outside organizations, including the NCMEC, the ICAC Task Force, Shift Wellness, and others.
A Gaggle Safety Team member will analyze the flagged piece of content to assess context and determine whether it is a false positive or if it might be something of concern. They will do this by looking at the text through the eyes of human intelligence and reading the surrounding text beyond the sentence in question.
If it is determined that the content in question indicates a mental health issue, a threat, or potential harm to a student, it is escalated to a second Gaggle Safety Team member who will look more carefully at the item and full context, categorize the concern, and determine the severity and urgency.
How does Gaggle process student content in non-English languages?
Gaggle uses a direct integration with a translation service in our interface, allowing our Gaggle Safety Team to translate content immediately when making a decision about student safety and well-being.
How does Gaggle interpret pictures?
Gaggle’s proprietary anti-pornography scanner (APS) offers unique real-time image analysis that detects and prevents pornographic images. The APS scans embedded and attached images as well as images in attachments like Word and PowerPoint documents. Even URL links to other websites are scanned for offensive content.
Does Gaggle monitor students’ social media accounts?
Gaggle does not monitor students' social media accounts, personal email accounts, personal devices, or web browsers. It only monitors content and activity, such as documents or chat messages, produced using a school-owned device, email address, or online tools within Google Workspace for Education, Microsoft 365, Google Chat, Microsoft Teams, and the Canvas learning management system.
Does Gaggle turn over data to law enforcement?
Most of the content identified by our technology is ultimately benign and does not require any engagement from the school emergency contact. In about 0.0036% of cases, our team believes that the content is sufficiently worrisome that an adult needs to be made aware and these items are labeled as incidents. In very, very rare instances—approximately 0.41% of these incidents, and only if we cannot reach the school-designated emergency contact after multiple attempts and believe a child is in serious crisis—we will contact the local authorities for a wellness check.
Gaggle only alerts law enforcement for an immediate wellness check when information indicates that a child’s life is in imminent danger and district-appointed emergency contacts are inaccessible. We know engaging law enforcement in any student situation is a last- and worst-case scenario, and take every step possible to avoid it unless a child’s life hangs in the balance.
