The head of Norwegian company Aiba says their system can monitor voice chat and flag up incidences of grooming, bullying, and radicalisation in near real-time
An AI-powered tool named Amanda, developed by Norwegian company Aiba, is set to detect signs of grooming in voice chat in real-time, allowing moderators to intervene promptly. The tool is integrated into the Odin voice chat platform by German firm 4Players. Hege Tokerud, CEO of Aiba, mentions that Amanda can identify various forms of social manipulation, including grooming, bullying, and radicalization.
The technology, inspired by the tragic story of Amanda Todd, has been trained on police records of cybergrooming cases, enabling early detection of harmful patterns in online conversations. Amanda analyzes context and language nuances, with results shared with human moderators for further assessment. The system supports nearly 40 languages and ensures GDPR compliance by encrypting voice data and storing it only when harmful speech is detected.
Aiba's anti-toxicity system has garnered interest from various companies seeking solutions for online safety. Tokerud notes a growing market demand for such tools, emphasizing the importance of creating a positive online community environment. While Aiba collaborates with multiple potential clients, they have an ongoing partnership with MovieStarPlanet, a Danish free-to-play online social game.
Other tech giants like Snap, Meta, and Roblox are also enhancing safety measures to protect children from online threats. Tokerud highlights that the AI system's ability to monitor voice chat for toxicity is not a future aspiration but a present reality, offering a solution to combat online toxicity effectively.
