February 19, 2021, Prince
4 stages of the evolution of YouTube and the role of AI in its development!!!
Do you know that, on average, YouTube records more than 2 billion monthly logged-in users?
Sounds fascinating, isn’t it?
But remember that with power comes its responsibility. YouTube influencers nowadays can reach out to millions of people in just a snap of a finger; any viral video on YouTube can pose a grave danger to any socio-economic community of a third-world country. How is YouTube tackling all this? And not just this; how does YouTube decide which video to promote and which video to flag down?
The answer is simple, AI.
Here are the four stages of the evolution of YouTube:
This phase of YouTube lasted for around seven years till 2012. This phase of YouTube was based on the view count. The use of AI in this phase was minimal. Google, the parent company of YouTube, was still in the initial phase of its experiment with AI. Google’s interest in AI started in 2011 with the project Google Brain which was restricted to recognize images through neural networks. In this phase of YouTube, videos were solely judged based on their views. This led to a massive loss in customer experience, as the genesis of clickbait can be found here. People were tricked by using catchy thumbnails to open their video, and YouTube used to promote that video without judging if the user has even seen the video or not.
This phase of YouTube expands from 2012 to 2016. In this phase, YouTube solved its drawback of clickbait videos destroying their user experience. They introduced “watch minutes” as a criterion for pushing videos rather than just the number of views. It solved all the clickbait issues as now, even if a video had a buttload number of views with negligible watch duration, the YouTube algorithm didn’t push the video to more people.
Finally, Machine learning is here
The third phase of YouTube was introduced in 2016. This phase came with a white paper released by YouTube. This white paper used many fancy terms filled with tech jargon but wasn’t a comprehensive approach to explain YouTube’s algorithm. But one thing was clear: YouTube was now going to inculcate AI and machine learning in its algorithms to make its user-experience easier and more relatable.
You can read the white-paper released by YouTube here.
Now that YouTube has become this powerful, it arises some problems too. Currently, YouTube is continuously working day-night to make its AI more and more powerful to detect obscene, vulgar, and offensive videos to take them down immediately. Now YouTube algorithm can easily detect these kinds of videos which saves them a lot of manpower. In fact, the head of EMEA, Cecile Frot-Coutaz, once said that YouTube’s “number one priority” is making sure YouTube protects its users from harmful content.
What we can learn from YouTube is that how we have to constantly adapt our business model and keep experimenting with new technologies to keep our business alive and in profit. YouTube adapted itself to inculcate AI and machine learning in its core functioning to provide its user a world-class experience.
YouTube has already realized that there is no future without AI. When will you realize it?
Hope to see your business using AI soon; till then,
Stay Safe and Keep Smiling!!!