Algorithmic Persuasion in the Digital Society
Lectures
Lecturers: Sophie Boerman and Brahim Zarouali
Semester 2 Block 2
,Lecture 1: Introduction to algorithms and the digital society
Defining algorithms
Algorithms: Encoded procedures for transforming input data into a desired output, based on specific
calculations.
• Comparable to a cake recipe
Algorithmic power: four main functions of algorithms
Algorithms can...
• Prioritization: making an order list
o Emphasize or bring attention to certain things at the expense of others (e.g., Google
Page rank)
o Make a list of the most important things while disregarding other things
• Classification: algorithms can easily pick a category
o Categorize a particular entity to given class by looking at any number of that entity’s
features (e.g., inappropriate YouTube content)
o Classify based on two categories (good to go or not good to go)
• Association: finding links
o Association decisions mark relationships between entities (e.g., OKCupid dating match)
o Finding certain patterns or links between two different people and partnering people up
• Filtering: isolating what’s important
o Including or excluding information according to various rules or criteria. Inputs to
filtering algorithms often take prioritizing, classification or association decisions into
account (Facebook news feed)
(Simplistic) Rule-based algorithms
Based on a set of rules you need to identify in advance
• If – Then statements (if person likes this content, give them more of this content)
o Advantage: easy to follow and formulate
o Disadvantage: only applicable to the specified conditions
Machine learning algorithms
Algorithms that “learn” by themselves (based on statistical models rather than deterministic rules)
• Trained based on a corpus of data which they use to learn how to make decisions without
human oversight
o Advantage: flexible and amenable to adaptations
o Disadvantage: need to be trained in order to perform well and black box (at some point
the decisions made by the algorithms won’t be understood)
Example of machine-learning algorithm: Facebook’s Deep Face algorithm
• Facial recognition system that identifies human faces in digital images
• Trained on a large identity label data set of four million facial images
, • Has 97% accuracy
• Not always being used for good reasons, Facebook sells this system to authoritarian regimes for
them to monitor people
• Important to ask, what are these algorithms being used for?
Recommender systems: important kinds of algorithms we see online
Provide suggestions for content that is most likely of interest to a particular user
• Decide which content to display to whom based on certain criteria
• Users are thus receiving distinct streams of online content
• Examples: FB news feed, Netflix movies, Spotify songs
Reasons why recommender systems are used:
1. Prevent choice overload: reduce pressure on users to choose something
2. Maximize user relevance: make sure content aligns best with user’s interests
3. Increase work efficiency: if you let the system recommend certain things, it is not necessary for
people to figure out what users want
Techniques
• Content based filtering: learn to recommend items that are similar to the ones users liked in the
past (based on similarity of items)
o If you like horror movies, Netflix will recommend you more of them
• Collaborative filtering: suggest recommendations to users based on items other users with
similar interests liked in the past
o People with the same gender, age range, people in your location
o Used very often when you start using a platform since it doesn't have enough data on
you to do content-based filtering
• Hybrid filtering: combine features from content based and collaborative systems while adding
other elements (e.g., demographics)
o Most used on most platforms
How do people perceive algorithms
Range of perspectives on algorithms: Appreciation <--> Aversion
• Algorithmic appreciation: people prefer algorithmic decision over human decision (even if they
know that the algorithm is inferior in quality to their human decision)
o Blind faith in automation: tourists trust Maps too much that they bike through a tunnel
meant for cars
▪ Blindly following technology without using your own human decision
• Algorithmic aversion: prefer human decision over algorithmic decision
Aversion or appreciation: a nuanced view
Depends on many other factors
• Type of task (Lee, 2018)
, • Level of subjectivity in decisions (Logg et al., 2019): how subjective is the task that needs to be
done?
• Individual characteristics (Araujo et al., 2020)
Three aims of the paper:
• Defined the concept of algorithmic persuasion
• Proposed the Algorithmic Persuasion Framework to understand the role of algorithms in
persuasive communication
• Presented a future research agenda based on the insight derived from the framework
Algorithmic persuasion: any deliberate attempt by a persuader to influence the beliefs, attitudes and
behaviors of people through online communication that is mediated by algorithms
• Persuasion is intentional and deliberate
• A persuader can be a person, brand, political and commercial actor
• Influence is important, impact should be generated through using algorithms, whether that be
weak/strong, short term/long term, it varies
• Through online communication: algorithms always occur online through media, web
• Mediated by algorithms: communication should be driven by algorithmic technology
Algorithmic Persuasion Framework is circular
• Accounts for feedback loops: new data we provide will always be used as new input, so the cycle
continues
• Model was based on two other models
Input: all data we feed into the algorithm
• Different distinctions: first, second, third party data
o First party: data collected by the company itself
o Second party: data collected from a collaborative second party (e.g., using data from
Google)
o Third party: data from external party specialized in data brokerage who sells you the
data to add to your data
• Implicit and explicit data:
o Explicit data: all data we explicitly leave behind when we go online (including personal
information we post to Facebook), we are aware of this
o Implicit data: data outside of our consciousness (IP address, search history)
Algorithm: putting data into a system to work
• Techniques: different algorithmic techniques (no one algorithm that does the trick, differences
that can solve specific problems)
• Objective of persuader: sender of algorithm can use it to change your attitudes, thoughts,
beliefs, behaviors
o Brands want you to have a more positive view of them
Lectures
Lecturers: Sophie Boerman and Brahim Zarouali
Semester 2 Block 2
,Lecture 1: Introduction to algorithms and the digital society
Defining algorithms
Algorithms: Encoded procedures for transforming input data into a desired output, based on specific
calculations.
• Comparable to a cake recipe
Algorithmic power: four main functions of algorithms
Algorithms can...
• Prioritization: making an order list
o Emphasize or bring attention to certain things at the expense of others (e.g., Google
Page rank)
o Make a list of the most important things while disregarding other things
• Classification: algorithms can easily pick a category
o Categorize a particular entity to given class by looking at any number of that entity’s
features (e.g., inappropriate YouTube content)
o Classify based on two categories (good to go or not good to go)
• Association: finding links
o Association decisions mark relationships between entities (e.g., OKCupid dating match)
o Finding certain patterns or links between two different people and partnering people up
• Filtering: isolating what’s important
o Including or excluding information according to various rules or criteria. Inputs to
filtering algorithms often take prioritizing, classification or association decisions into
account (Facebook news feed)
(Simplistic) Rule-based algorithms
Based on a set of rules you need to identify in advance
• If – Then statements (if person likes this content, give them more of this content)
o Advantage: easy to follow and formulate
o Disadvantage: only applicable to the specified conditions
Machine learning algorithms
Algorithms that “learn” by themselves (based on statistical models rather than deterministic rules)
• Trained based on a corpus of data which they use to learn how to make decisions without
human oversight
o Advantage: flexible and amenable to adaptations
o Disadvantage: need to be trained in order to perform well and black box (at some point
the decisions made by the algorithms won’t be understood)
Example of machine-learning algorithm: Facebook’s Deep Face algorithm
• Facial recognition system that identifies human faces in digital images
• Trained on a large identity label data set of four million facial images
, • Has 97% accuracy
• Not always being used for good reasons, Facebook sells this system to authoritarian regimes for
them to monitor people
• Important to ask, what are these algorithms being used for?
Recommender systems: important kinds of algorithms we see online
Provide suggestions for content that is most likely of interest to a particular user
• Decide which content to display to whom based on certain criteria
• Users are thus receiving distinct streams of online content
• Examples: FB news feed, Netflix movies, Spotify songs
Reasons why recommender systems are used:
1. Prevent choice overload: reduce pressure on users to choose something
2. Maximize user relevance: make sure content aligns best with user’s interests
3. Increase work efficiency: if you let the system recommend certain things, it is not necessary for
people to figure out what users want
Techniques
• Content based filtering: learn to recommend items that are similar to the ones users liked in the
past (based on similarity of items)
o If you like horror movies, Netflix will recommend you more of them
• Collaborative filtering: suggest recommendations to users based on items other users with
similar interests liked in the past
o People with the same gender, age range, people in your location
o Used very often when you start using a platform since it doesn't have enough data on
you to do content-based filtering
• Hybrid filtering: combine features from content based and collaborative systems while adding
other elements (e.g., demographics)
o Most used on most platforms
How do people perceive algorithms
Range of perspectives on algorithms: Appreciation <--> Aversion
• Algorithmic appreciation: people prefer algorithmic decision over human decision (even if they
know that the algorithm is inferior in quality to their human decision)
o Blind faith in automation: tourists trust Maps too much that they bike through a tunnel
meant for cars
▪ Blindly following technology without using your own human decision
• Algorithmic aversion: prefer human decision over algorithmic decision
Aversion or appreciation: a nuanced view
Depends on many other factors
• Type of task (Lee, 2018)
, • Level of subjectivity in decisions (Logg et al., 2019): how subjective is the task that needs to be
done?
• Individual characteristics (Araujo et al., 2020)
Three aims of the paper:
• Defined the concept of algorithmic persuasion
• Proposed the Algorithmic Persuasion Framework to understand the role of algorithms in
persuasive communication
• Presented a future research agenda based on the insight derived from the framework
Algorithmic persuasion: any deliberate attempt by a persuader to influence the beliefs, attitudes and
behaviors of people through online communication that is mediated by algorithms
• Persuasion is intentional and deliberate
• A persuader can be a person, brand, political and commercial actor
• Influence is important, impact should be generated through using algorithms, whether that be
weak/strong, short term/long term, it varies
• Through online communication: algorithms always occur online through media, web
• Mediated by algorithms: communication should be driven by algorithmic technology
Algorithmic Persuasion Framework is circular
• Accounts for feedback loops: new data we provide will always be used as new input, so the cycle
continues
• Model was based on two other models
Input: all data we feed into the algorithm
• Different distinctions: first, second, third party data
o First party: data collected by the company itself
o Second party: data collected from a collaborative second party (e.g., using data from
Google)
o Third party: data from external party specialized in data brokerage who sells you the
data to add to your data
• Implicit and explicit data:
o Explicit data: all data we explicitly leave behind when we go online (including personal
information we post to Facebook), we are aware of this
o Implicit data: data outside of our consciousness (IP address, search history)
Algorithm: putting data into a system to work
• Techniques: different algorithmic techniques (no one algorithm that does the trick, differences
that can solve specific problems)
• Objective of persuader: sender of algorithm can use it to change your attitudes, thoughts,
beliefs, behaviors
o Brands want you to have a more positive view of them
