Detox Browser

Prevent infoxication, control what you read on the web
-An Initiative by RISHA Lab IIT Tirupati

About

With Covid-19 continuing to rage across the country, it has begun to take a serious toll on the mental health of senior citizens and sensitive individuals due to information overload. Experts have suggested a reduction in screen time and filtering out misinformation among other things. DetoxBrowser enables a user to filter out content which he/she might feel is adversely affecting their mental well-being. In addition to default filters, the extension enables a user to fine-tune the browsing experience as per their individual requirements. The extension analyzes sentiment, categories and blacklisted topics to prevent the user from being overwhelmed with information that can possibly traumatize the user. The extension achieves this through simple keyword-based checks followed by detailed analysis. Even though it is mainly aimed at google search results it also checks websites for topics blacklisted by the user. The extension is in open beta and its features are being continuously improved.

Built by Noble Saji Mathews

Under the guidance of Dr. Sridhar Chimalakonda

Getting Started

1. Setup

After installation, Tap on the extension from the tray . This will show a popup to control the filter sensitivity. Tap on the wrench icon in the popup to open options panel.

2. Personalize

Polarity | Eg. -5: Death (makes the extension more sensitive in Google Search)

Filter | Eg. Gun violence (Keyword doesn't appear even out of search)

+ options shown in image

P.s. The extension is meant to be personalised and improve over time based on your filters, this means that it tends to flag based on keywords alone pretty easily unless tuned to personal preferences. However, for future builds we plan to replace this trigger happy behaviour with a secondary check and have an easier slider system for which we would appreciate sharing your personalised polarity lists and blacklists.
If you are comfortable with this, do send the same through this form to enable us to keep improving the user experience.

Features

  • Temporarily stop getting search results about information that may be sensitive to the user. For example, one could add words such as “Covid” and others like “Stocks”, “News about Abuse”, “Murders / Deaths”, “Adult / Slang” etc.

  • Uses a basic but quick wordlist based sentiment analysis to flag badly worded or concerning results.

    • Can add in words and tweak the sensitivity of filter based on issues.

    • Shows tags the search result is about on hover, and the source of the result so can be unhidden if the user wants to.

    • Note: Search results are not analysed on special pages such as that obtained when directly searching for covid and or other topics with misinformation surrounding them to prevent these from being blocked.

  • Abusive language filtered with leet and profanity checks.

  • Option between direct removal from google search if blacklisted keywords detected.

    • Blacklist supports phrases as well

  • Remote server for categorization and detailed analysis of flagged articles.

    • Categorization based on 20 years Times of India headlines dataset

  • Onload website blocking if blacklisted words detected on any non -google search based site.

Upcoming Features :

  • Adding a secondary check with context-aware sentiment analysis to reduce false positives without user intervention.

  • Changing / Retraining the categorization ML model to a more accurate variant.

  • Starter keywords to slide default levels for a casual user.

  • Taking into account dislikes and image metadata for more features surrounding media.

  • Extend native full-fledged support for social apps.