Project Apate
Faink No More ( don't utter anything Apate will check what you say ). Apate tags the video's text to categories like, truth, lie, exaggeration, hate speech etc. Tags are crowdsourced like wikipedia.
Created on 14th October 2018
•
Project Apate
Faink No More ( don't utter anything Apate will check what you say ). Apate tags the video's text to categories like, truth, lie, exaggeration, hate speech etc. Tags are crowdsourced like wikipedia.
The problem Project Apate solves
59 out of 100 people don’t even read the article before sharing it, and others are too lazy to actually verify the correctness of it, so essentially people are busy spreading any content - we wish to change that.
The internet is a great tool, where any thought can reach to millions of people in no time.
But this great tool needs to be handled carefully. It has the power to manipulate people, it can ignite rage and spread hatred.
This coupled with the blind content sharing, spreads the negative emotions.
The reason for hatred, false and exaggerated content generated by online content generators is because it is intended to attract more advertisers, and more advertisers wants more eyeballs, so the best interest of the content creators is used to add spice to the content. This leads to unrest in general public as they blindly follow the emotions.
Apate solves the following problems :-
-
It annotates the video content, by marking the part of the subtitles as truthful, lie, exaggerated, hatred etc. Thus it allows blind sharers to make informed decision before arriving at any conclusion.
-
Using collective knowledge to make informed decisions, Say I am watching some economic video, where someone is saying rupee will go down etc. I am not the subject matter expert and the only thing that I am getting from the video is fear. But I someone who has knowledge about that can annotate the text as Truthful, Lie, exaggerated. I can be more certain about the video.
-
Regulating online content: It will create the ultimate repository of the video, highlighting which part of the video are truth, lie, exaggerated etc.
Challenges we ran into
Challenges we ran into :-
- Getting the subtitles of the video
- Parsing the subtitles, cleaning them, reading them etc because of multiple ways YouTube subtitles are generated
- Developing a robust frontend
- Figuring out and creating a user experience for the unique flow