Chapter 3

Propagation of information and

misinformation

01/13
1
Overview
❖ Information - definition, information collection methods and use
❖ Nature of Information spread
❖ Digitization of information and its effects
❖ Misinformation propagation and its effects
❖ Examples on misinformation spread
❖ Guidelines to mitigate the spread of misinformation
❖ Roles of Computer Science to tackle misinformation

01/13
2
Information
Information is data that has been processed to make it meaningful and
useful.

Data vs Information
• Data is raw and unorganized
• Information is organized, accurate
form of data. It is interpreted within
a useful context

01/13
3
Different methods of information
collection
● Useful for non quantifiable information.
01 Observations ● Can be influenced by personal bias.

● Information is gained through a control over

variables.
02 Experiments ● The quality of information depends on the design of
the experiments.
● Collects information from interested people on a
03 Surveys specific aspects. Usually done through
questionnaires.

Subjective ● If objective methods cannot be used, subjective

04 Estimations opinion is gathered from experts.
01/13
4
Different methods of information
collection
● Information is gathered on individual basis.
05 Interviews ● Can be time consuming and information has to be
processed before using

● Information can be collected from publications,

Government agencies, social media, transaction
06 Existing sources records, different organizations, and many other
sources.

01/13
5
Use of information
Market analysis
• Target audience cluster
• People’s preference
• Current trend
• Future trend analysis

Political opinion formation

• Information gathering for campaign
• Understanding public preference
• Influencing public preference
01/13
6
Use of information
Research
• Healthcare
• Share market
• Computer vision

In war and conflicts

• Encryption/Decryption of radio
information
• Raising awareness
01/13
 Nature of information spread
A huge number of people are using various social media websites
• Twitter had 330 million monthly active users as of 2019
• Facebook had 2.50 billion monthly active users as of 2019
• YouTube had 2 billion monthly active users as of 2019
Most of these users are connected with each other, these connections can be
presented with a graph

• Nodes are the people within a social network

• Edges represents the connection between them (People in Facebook friend list,
Followers in twitter)
01/13
 Types of connections
• Bi-directional connection
• Information spreads from user A to user B and vice versa, as both user can see the
information of each other.
• For example, people in Facebook friend list

• Unidirectional connection
• Information only spreads from user A to user B or vice versa, as only one user can see the
information of other
• For example, someone following a public figure in twitter

• Although, if two people are following each other, the connection will be bi-directional
01/13
 Spread of information

• Information does not stay contained within a single social media website.
• People who use multiple social media spreads information from one
social media to another.
• These people are sometimes referred as ‘Bridge nodes’.
• Therefore, the actual social network is a massive interconnected graph
consisting for users from all the different social media websites.
01/13
Information spread in social network

01/13
11
 Think about it…
Abid (A) and Rohan (R) are two friends connected on Facebook. Millie (M), another
Facebook member, have both Abid and Rohan in her friend list, but unfollowed Rohan
due to his excessive meme sharing. Rohan also has a twitter account where he follows a
celebrity called Jack (J). Rohan and his friend Sharma (S) follows each other on twitter.
Additionally, Rohan follows Millie’s Instagram account. Now, try to draw an information
propagation graph based on the information.

01/13
 Think about it…
Abid (A) and Rohan (R) are two friends connected on Facebook. Millie (M), another
Facebook member, have both Abid and Rohan in her friend list, but unfollowed Rohan
due to his excessive meme sharing. Rohan also has a twitter account where he follows a
celebrity called Jack (J). Rohan and his friend Sharma (S) follows each other on twitter.
Additionally, Sharma follows Millie’s Instagram account. Now, try to draw an information
propagation graph based on the information.

01/13
 Think about it…
Abid (A) and Rohan (R) are two friends connected on Facebook. Millie (M), another
Facebook member, have both Abid and Rohan in her friend list, but unfollowed Rohan
due to his excessive meme sharing. Rohan also has a twitter account where he follows a
celebrity called Jack (J). Rohan and his friend Sharma (S) follows each other on twitter.
Additionally, Sharma follows Millie’s Instagram account. Now, try to draw an information
propagation graph based on the information.

01/13
 Think about it…
Abid (A) and Rohan (R) are two friends connected on Facebook. Millie (M), another
Facebook member, have both Abid and Rohan in her friend list, but unfollowed Rohan
due to his excessive meme sharing. Rohan also has a twitter account where he follows a
celebrity called Jack (J). Rohan and his friend Sharma (S) follows each other on twitter.
Additionally, Sharma follows Millie’s Instagram account. Now, try to draw an information
propagation graph based on the information.

01/13
 Think about it…
Abid (A) and Rohan (R) are two friends connected on Facebook. Millie (M), another
Facebook member, have both Abid and Rohan in her friend list, but unfollowed Rohan
due to his excessive meme sharing. Rohan also has a twitter account where he follows a
celebrity called Jack (J). Rohan and his friend Sharma (S) follows each other on twitter.
Additionally, Sharma follows Millie’s Instagram account. Now, try to draw an information
propagation graph based on the information.

• Can you tell how many ways Sharma

can see Jack’s public information
despite not being connected to him
directly?
• If you unfollow someone on social
media, does the information
propagation from him stop completely?
Explain why or why not based on the
given graph.

01/13
 Think about it… (Optional)
● Let’s consider that the graph in the previous slide completely isolated (This users
are not connected with any other user). Now, you have to choose one user to
advertise a product that you are selling. Theoretically speaking, if you need the
advertisement to reach maximum number of people in that social network, whom
would you choose? Why?

● Now in real case scenario, things are much more complex. The network graphs are
gigantic and many external factors come into play (For example, if you like/react
someone’s post more, you usually get more exposed to their updates. Also, many
people may not have any preference to your product). It’s not possible to find
solutions with bare eyes, but an automated data driven analysis may find out the
solution. Computer science discusses these data driven approaches.

01/13
Digitization of information

• Conversion of text, pictures, or sound into a

digital form is the digitization of information

• Requires much less space to store and can be

stored in great volumes

• May spread much faster as information can be

shared instantly
01/13
Digitization of information
• Stealing digital information is easier
• Can be modified and fabricated fairly easily

01/13
Misinformation
• Definition - Information that is false or deceitful and shared without
the knowledge of it being false
• Types of misinformation
• Fake news
• Rumors
• Information without proper context
• Satire
• Modified/misleading content

01/13
Misinformation spreads faster
• Looking at 126,000 stories sent by almost 3 million people, study
found that fake news spread much faster than regular information

• A fake news has 70% more change to earn a retweet.

• Reasons:
• Fake news is usually more dramatic than real news
• Fake news is designed to attract more attention

01/13
Recent incidents of misinformation
propagation around us
•Year 2019-2020 had the highest rate of
fake news propagation

•The number of requests to check

veracity increased by 20% in Bengal-
Indian region in 2020

•Fake news spread like wildfire during

covid-19

01/13
Recent incidents of misinformation
propagation around us

01/13
 Misinformation spread example

Someone on twitter spreading

misinformation on Covid-19
antivirus.
• People retweet the post
without judging any
authenticity or validity
of the post.
• Causes panic among
crowds as the news
spread
01/13
24
Misinformation spread example
Twitter dealt with
these misleading
posts by introducing
label for Tweets
containing synthetic
and manipulated
media.
• Warns the users
about false claims.

01/13
25
More examples of misinformation on social
networks

01/13
26
Even though social sites try to control the
spread of misinformation,

Is it always successful?

01/13
27
How to mitigate the spread of
misinformation
1. Be critical when you look at social media.
2. Don’t leave false information in your online networks.
You can politely ask the person who shared it to remove it.
3. Report the false information to the platform
administrators.
4. When in doubt, take the time to verify the shared
information.
5. Make more noise than people who share false
information.

01/13
28
 Role of computer science to prevent
misinformation spreading
•Various Artificial Intelligence (AI) driven approach to detect misinformation

•Automated approaches to make the misinformation less visible to users

•Works the same way spam detection works

•With more data being generated, AI driven approaches are improving over time

•AI does not only detect misinformation in texts, it can also detect image
manipulation (Deep fake detection challenge)
01/13
 Role of computer science to prevent
misinformation spreading
•Grover - An AI witch learned to write fake news in order to detect it
misinformation

•Various other methods used Neural Network relate techniques to detect fake
information

•However, due to sheer variation of misinformation, these AI systems are still not
perfect

•A combination of AI and human awareness development may significantly

reduce misinformation propagation 01/13
Thank you

01/13