of daily social media interactions monitored for abusive content in near real time
real-time feedback loop continuously retrains models to keep pace with evolving children's communication styles
independent pipelines detect abusive language, emojis, and media across content types simultaneously
As social media use continues to grow, children are increasingly exposed to potentially harmful content. OneSix partnered with a fast-growing technology company to address this challenge, aiming to monitor and detect abusive content effectively while respecting the unique style and nuances of children’s online communication. This required developing a solution that could handle the vast scale of daily content, including millions of interactions, and adapt to the evolving ways children communicate online.
OneSix built a custom, end-to-end platform with multiple machine learning models specifically trained to recognize abusive language, emojis, and media in children’s online communication. The solution included independent autoscaling pipelines for triaging content types, analyzing web and media content, and processing natural language text to detect abuse as it occurs.
To adapt to the unique and changing nature of children’s online language, the platform included a real-time feedback loop for continuous model updates. This loop enabled strategic selection of content for expert annotation, feeding those insights back into model training to keep pace with evolving trends. Additionally, a suite of performance monitoring tools was developed to track model accuracy and responsiveness, ensuring both effective monitoring and high-quality data for ongoing improvements.
The solution has enabled near real-time identification of abusive content across millions of daily social media interactions, providing a powerful safeguard against cyberbullying and harmful language exposure. The adaptive model and feedback loop ensure that the platform remains effective as communication styles shift, delivering timely and accurate detection to keep children safe online. The collaboration with OneSix has equipped the client with a robust, scalable system that continuously learns and improves, meeting the demands of online safety in an ever-changing digital landscape.
More to explore
Through a three-phase approach, including a comprehensive assessment, tailored recommendations, and a clear roadmap, OneSix helped the bank set the foundation for improved operational efficiency, enhanced customer experiences, and future-ready data strategies.
.png)
OneSix built an edge-based AI computer vision system that detects shrink-wrap packaging defects in real time and scales across production lines from a single multi-task model, removing per-line calibration and operator subjectivity from the inspection process.
OneSix implemented a data warehouse solution for a commercial printing client, transforming IoT sensor data into real-time, actionable insights that enhanced operational visibility and reduced downtime.