acrofan

Industry Economy TECH GAME
Society Comfort AUTO MEDIA

Artificial intelligence monitoring : “must-have” in an online world

  • Wednesday, February 3, 2021, 8:00 am
  • ACROFAN=Jae-Yong Ryu
  • jaeyong.ryu@acrofan.com
Hyperconnect has excellent staff working for AI monitoring capabilities.

COVID-19’s impact on professional and commercial spheres — from education to entertainment, distribution to shopping, healthcare to culture, and far beyond — have been well-documented. Perhaps less well-understood, however, is the long-term influence COVID will have on our social tendencies and mental health. Few people would argue, however, that more and more people are connecting, interacting and forming relationships online.

- Online safety measures are needed for online lives

As social platform usage intensifies, so too can online misbehavior. Facebook, the world's largest social media platform, approaches the issue using artificial intelligence technology. For example, all post comments are checked by AI-powered tools and automatically hidden if deemed offensive or in violation of user policy. If a user’s proposed comment is identified as similar to previously-flagged comments, ‘post confirmation’ warning messages are automatically shown to the posting user. When Facebook’s AI is unable to fully confirm if a post is offensive, potentially-flagrant content is forwarded to human monitoring staff for further review.

Facebook is not alone in its proactive approach. YouTube, the world's largest video platform, too employs machine learning tools to detect violent and other content that may threaten child safety. YouTube’s in-house Trust & Safety Team deploys tools to constantly monitor content, rapidly applying age restrictions on videos that are deemed not suitable for younger users. Amazon’s Twitch, the world's largest game streaming platform, also utilizes AI to automatically identify offensive comments and deliver them to monitoring personnel for review. After inspections are completed, feedback is provided to other users.

- Hyperconnect is among the world’s best for AI monitoring capabilities

Hyperconnect, the Seoul-based technology company behind real-time video chat app Azar, which has exceeded 545 million cumulative downloads globally, deploys a comprehensive system of AI-enabled tools to ensure high standards of community safety. For its advanced “on-device” AI, Hyperconnect’s deep learning-based visual detection models can now pre-block content within 0.006 seconds depending on device specification. Given the system’s operation on device, privacy is ensured as video is not routed up through a server. Hyperconnect also deploys a comprehensive AI system to pre-screen user-submitted reports, remedying most user-flagged issues without needing human agent intervention.

“Amongst the non-gaming apps market globally, social discovery apps are taking the top spots. As COVID-19 accelerates the transition to a 'digital daily life', global social platforms should spare no expense in ensuring a healthy and safe digital environment for users. Substantial investment in automatic monitoring systems is an obvious starting point” according to a representative from Hyperconnect.