Video content forms a sizeable percent of online data today. Herein live video streaming is a growing market. Live streaming is becoming increasingly popular across domains in relaying events live to end users. Over time the quality of the live stream has also improved drastically. Not just from a resolution standpoint but also from an audio-video synchronization standpoint. What next for the world of live streaming? While evolutions in the space of video quality will and need to continue to happen. The one important area to focus is searching through the live video clips – this needs to happen for two reasons:
- To make the content discoverable. Only when the data becomes searchable does it become discoverable. Otherwise. for now only meta data about the content is indexable. searchable and discoverable.
- Secondly. more importantly. since live video clips is a very important tool in the hands of the global population at large. It has to become searchable to ensure no objectionable content is being shown. This is important both from regulators standpoint as well as the social platform/video tool standpoint. In the past. we have had some projects where we have had to view videos to ensure there is no objectionable. Malicious content and if so flag them to be blocked. While this was a manual task to take up. Today’s global video consumption makes this task very challenging on various grounds –
- Volume of video content. Specifically live video is very high and daunting that it does not warrant manual reviewing and flagging.
- What is considered objectionable at a global level is too high to keep track of through a manual checklist. Let’s say. There is an objectionable violent act that is streamed live in Brazil. How would a tester in Asia be able to say whether it is objectionable or an accepted form of say martial art in the country?
Given all these nuances. An automated way seems the best solution to monitor streamed content live. add tags. check for whether it is safe for viewing or needs to be pulled down. The automated solution gaining popularity to solve these problems is artificial intelligence. Leading social platforms such as Facebook and Twitter have already been investing in this area to leverage artificial intelligence to monitor their live feeds and video clips. With recent terror attacks that are suspected to have been live streamed. This process of reviews has further been strengthened in the recent days. Machine learning algorithms can further add to the richness of this search and review process. Making it a robust tool though not fool proof in the coming years. Exciting that this large piece of online content is coming into a controlled radar which will make live streaming more secure. And relevant for the global masses.