CoTruth

Meet CoTruth AI Assistant which stops the spread of misinformation related to covid 19 on the social media platforms and keeps mental health in check. CoTruth makes fact checking rewarding and fun.

Comment

Inspiration

One of the key challenges of the pandemic has been misinformation and disinformation about the COVID-19 pandemic. This has led to a variety of problems, including stigma from the disease, failing to take appropriate measures, and getting the right treatment.

How do we enable seamless flow of trustworthy news and scientific facts to the broader population?

How do we equip citizens with the right information to combat the constant, viral stream of distrustful information from social media and other sources?

How can we encourage people to take appropriate preventative measures, to seek proper treatment and avoid excessive or incorrect use of unapproved medicines and treatment?

What can we do to combat the stigma of COVID-19 that is created due to unbalanced and sensationalized news?

There was an incident during pandemic where workers gathered at Bandra railway station who wanted to go back to their hometown as they were stuck during pandemic and misinformation of trains starting caused chaos ? What if Higher Authorities could know such misinformation which can cause harm and chaos in society

Public gatherings are especially dangerous because while people attempt to discuss COVID, they are more likely to acquire the disease through the large group.

link

Misled Science Scientific development could be misled by publishing false reports and you could lose years and years worth of research due to one error in the findings.

False Cures False cures could be potentially unhealthy for the patient, and if the patient believes the cure was effective, they could risk infecting others due to unawareness.

The problem that many people face is the abundance of misinformation and false reports from numerous sources. I believe that it is very important to have access to accurate information as efficiently as possible so day-to-day people can prepare and protect themselves from false news reports. Instead of trying to get rid of misinformation entirely (since it is a very difficult goal to accomplish) we should provide a source of constant reliability and only publish scientifically proven statements.

During a seminar Shraddha Ma'am from (mystory.com) spoke about how misinformation spreads and even if the sources can Target misinformation they are failing . And the reason is its the people's not the source "We should make it a habit to check facts" ma'am told

What it does

CoTruth is an AI assistant which can be installed in a mobile device unlike bots found on Telegram, Whatsapp and other platform CoTruth is platform independent One tap solution which wasn't platform dependent simple (bots are specific to platforms)

When CoTruth is triggered it checks the information which is present on the screen and scans for the message using OCR

CoTruth multilingual support with the ability to detect printed materials as well .

Okay so India has such 60 sources which already tackles the same problem then how our solution stands out .

Now, How do we make fact checking a habit

We came up with a rewarding system that gives users points which in turn they can use it as discount coupons, offers etc. We made checking facts like gamification and addiction . There is a daily task for checking forwarded messages for scans , reporting messages if any , talking to an assistant about mental health in turn completion of tasks would give them points.

As the points are gathered and daily strak of chichking for misinformation the user account levels up unlocking the titles and badges such as "Master Buster" etc on their profile.

"So how do we target rural areas ? Too even though the main source is urban and semi urban areas"

it has multilingual support and in rural areas people live in join family so even if a younger user has phone the information would spread through words (typical thing that happens in village )

The daily tasks also known as goals have some activities which targets mental health of individuals as well

How we built it

CoTruth uses reputable resources to provide answers to misleading information such as WHO and other sources.

CoTruth classifies the misinformation using bert sentiment analysis which decides whether the following message is fake or legitimate; the classification of misleading information is coded using python.

OCR is used to extract the texts from the screen shots whenever the assistant is triggered

For recognising the the speech and converting it into text CoTruth uses Google translate API which has multilingual support for identifying and converting the given speech input into a query which is further sent to bert model classifier

CoTruth will further use algorithms like GPT3 or GPT2 for generating a reply to the query and ocr to extract text in the form of query.

The front-end is made up with frameworks which are web-based progressive web based on react JS

Challenges we ran into

Identification of the sources of misinformation was a huge challenge but it becomes easier if the misinformation is reported at earlier stages which increases the possibility of tracing the source.

Stopping each and every misinformation to be spread after detection

Reducing the steps required for fact checking whenever there is a suspicious message.

To indulge people in fact checking before forwarding the messages on social platforms.

Accomplishments that we're proud of

CoTruth has made fact checking as simple as a click along with that the assistant is platform independent

Gamification feature of the AI assistant has made fact checking habit and rewarding

Could be used by any age group; a single click is enough.

Counters the misinformation as the misinformation when identified can be shared to authorities to take the actions

What we learned

Stopping the spread of misinformation is very crucial step towards well being of a society

Inducing a everyday habit which will reward user for doing a right work will result in moral boost

What's next for CoTruth

CoTruth is very flexible in terms of domain; it can be trained on any domain for even targeting any misinformation.

further the misinformation data can be stored in a database which later can be used to analyze and train data on the captured information.

Additionally the features of mental health could also be added to the AI like talking and relaxing exercises

Further in app rewards could be provided such as discount coupon subscriptions like in Google pay and steps-set-go

Built With

Share this project:

Updates