Algorithmic Persuasion in the Digital Society: Summary
Week 1: Introduction to algorithms and the digital society
Lecture 1
Algorithms: Encoded procedures for transforming input data into a desired output, based
on specified calculations
Algorithmic power (desired output):
● Prioritization (making an ordered list): Emphasize or bring attention to certain things
at the expense of others
● Classification (picking a category): Categorize a particular entity to given class by
looking at any numbers of that entity’s features
● Association (finding links): Association decisions mark relationships between entities
● Filtering (isolating what’s important): Including or excluding information according to
various rules or criteria. Inputs to filtering algorithms often take prioritizing,
classification, or association decisions into account.
Rule-based algorithms:
● Based on a set of rules or steps
● IF - THEN statements
● + Quick, easy to follow/understand
● - Only applicable to the specified conditions
Machine learning algorithms (supervised & unsupervised):
● Algorithms that learn by themselves: based on statistical models rather than
deterministic rules/set of specific instructions.
● These algorithms are trained based on a corpus of data from which they may learn to
make certain kinds of decisions without human oversight
● + Flexible and amenable to adaptations
, ● - Needs to be trained & black-box
ML-Based conversations & how they work:
1. Natural language processing (NLP): The chatbot analyzes the user’s message to
understand intent and entities
2. Machine learning models: The chatbot uses ML models trained on past
conversations and user data to predict the best response. It can learn from
experience so it improves over time.
3. Context awareness: Unlike rule-based bots that forget past interactions, ML chatbot
can remember context of a conversation
4. Personalization & Learning: ML chatbots can remember past orders and suggest
options based on user preference.
5. Handling variability in responses: ML chatbot understands variation in human speech
and recognizes them all as the same intent.
Logic: Train the algorithm on a sample of (labeled) data, and then it can be used for making
decisions about other unseen data.
Platforms use ML algorithms.
Types of ML algorithms:
● Supervised learning: The algorithm learns from labeled data, which means the input
data is paired with the correct output.
● Unsupervised learning: Deals with unlabeled data. The algorithm tries to find patterns
or structure in the data without any guidance on what to look for.
● Semi-supervised learning: Combination of supervised and unsupervised learning
● Reinforcement learning: Involves an agent learning to make decisions by interacting
with an environment
Recommender systems are algorithms that provide suggestions for content that is most
likely of interest to a particular user.
Rationale of recommender systems:
● Avoid choice overload
● To maximize user relevance
● To increase work efficiency
Recommender system techniques:
● Content-based filtering: Algorithms recommend items that are similar to the ones that
the user liked in the past
● Collaborative filtering: Algorithms suggest recommendations to the user based on
other users with similar taste
● Hybrid filtering: Algorithms combine both features with other additional elements
People perceive algorithms: Appreciation versus aversion
Algorithmic appreciation:
, ● People rely more on advice from algorithms than other people
● Despite blindness to the process: black-box
● Automation bias: Computers are objective, rational, neutral, reliable etc. Humans
tend to over-rely and trust information more from automation
Algorithmic aversion:
● Tendency to prefer human judgments over algorithmic decisions, even when human
decisions are inferior
● Less tolerance from errors from algorithms than from humans
● People are averse because they don’t understand the algorithmic process
● Algorithmic anxiety: Lack of control and uncertainty
Aversions or appreciation depend on many other factors:
● Level of subjectivity in decisions, such as subjective / objective domains
● Individual differences
● More research needed
Algorithmic awareness: Very low
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.
Algorithmic persuasion framework:
Article 1: The Algorithmic Persuasion Framework in online communication:
Conceptualization and a future research agenda (Boerman et al., 2022)
, An algorithm is a set of step-by-step instructions computers are programmed to follow to
accomplish certain tasks. They are often used in data-driven media landscapes to
automatically and dynamically modify the content users see.
Users are more influenced than ever by personalized and distinct streams of online content.
-> Algorithms are transforming online platforms into codified environments that expose users
to the content that is likely to be most persuasive to them.
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.
Deliberate attempt means that the attempt is purposefully initiated by the persuader. The
persuader can be a brand, organization or person.
Influence is a change in people’s beliefs, attitudes and behaviors after exposure to
algorithm-mediated communications.
This can vary from:
● Very weak to very strong
● Conscious or unconscious
● With or without consumers’ awareness
● Short term or long term
● Intended or unintended
Online communication is the transmission of algorithm-mediated communication from
senders to receivers online.
The algorithmic persuasion framework is seen as a dynamic process.
5 components of the algorithmic persuasion framework:
● Input
● Algorithm
Week 1: Introduction to algorithms and the digital society
Lecture 1
Algorithms: Encoded procedures for transforming input data into a desired output, based
on specified calculations
Algorithmic power (desired output):
● Prioritization (making an ordered list): Emphasize or bring attention to certain things
at the expense of others
● Classification (picking a category): Categorize a particular entity to given class by
looking at any numbers of that entity’s features
● Association (finding links): Association decisions mark relationships between entities
● Filtering (isolating what’s important): Including or excluding information according to
various rules or criteria. Inputs to filtering algorithms often take prioritizing,
classification, or association decisions into account.
Rule-based algorithms:
● Based on a set of rules or steps
● IF - THEN statements
● + Quick, easy to follow/understand
● - Only applicable to the specified conditions
Machine learning algorithms (supervised & unsupervised):
● Algorithms that learn by themselves: based on statistical models rather than
deterministic rules/set of specific instructions.
● These algorithms are trained based on a corpus of data from which they may learn to
make certain kinds of decisions without human oversight
● + Flexible and amenable to adaptations
, ● - Needs to be trained & black-box
ML-Based conversations & how they work:
1. Natural language processing (NLP): The chatbot analyzes the user’s message to
understand intent and entities
2. Machine learning models: The chatbot uses ML models trained on past
conversations and user data to predict the best response. It can learn from
experience so it improves over time.
3. Context awareness: Unlike rule-based bots that forget past interactions, ML chatbot
can remember context of a conversation
4. Personalization & Learning: ML chatbots can remember past orders and suggest
options based on user preference.
5. Handling variability in responses: ML chatbot understands variation in human speech
and recognizes them all as the same intent.
Logic: Train the algorithm on a sample of (labeled) data, and then it can be used for making
decisions about other unseen data.
Platforms use ML algorithms.
Types of ML algorithms:
● Supervised learning: The algorithm learns from labeled data, which means the input
data is paired with the correct output.
● Unsupervised learning: Deals with unlabeled data. The algorithm tries to find patterns
or structure in the data without any guidance on what to look for.
● Semi-supervised learning: Combination of supervised and unsupervised learning
● Reinforcement learning: Involves an agent learning to make decisions by interacting
with an environment
Recommender systems are algorithms that provide suggestions for content that is most
likely of interest to a particular user.
Rationale of recommender systems:
● Avoid choice overload
● To maximize user relevance
● To increase work efficiency
Recommender system techniques:
● Content-based filtering: Algorithms recommend items that are similar to the ones that
the user liked in the past
● Collaborative filtering: Algorithms suggest recommendations to the user based on
other users with similar taste
● Hybrid filtering: Algorithms combine both features with other additional elements
People perceive algorithms: Appreciation versus aversion
Algorithmic appreciation:
, ● People rely more on advice from algorithms than other people
● Despite blindness to the process: black-box
● Automation bias: Computers are objective, rational, neutral, reliable etc. Humans
tend to over-rely and trust information more from automation
Algorithmic aversion:
● Tendency to prefer human judgments over algorithmic decisions, even when human
decisions are inferior
● Less tolerance from errors from algorithms than from humans
● People are averse because they don’t understand the algorithmic process
● Algorithmic anxiety: Lack of control and uncertainty
Aversions or appreciation depend on many other factors:
● Level of subjectivity in decisions, such as subjective / objective domains
● Individual differences
● More research needed
Algorithmic awareness: Very low
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.
Algorithmic persuasion framework:
Article 1: The Algorithmic Persuasion Framework in online communication:
Conceptualization and a future research agenda (Boerman et al., 2022)
, An algorithm is a set of step-by-step instructions computers are programmed to follow to
accomplish certain tasks. They are often used in data-driven media landscapes to
automatically and dynamically modify the content users see.
Users are more influenced than ever by personalized and distinct streams of online content.
-> Algorithms are transforming online platforms into codified environments that expose users
to the content that is likely to be most persuasive to them.
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.
Deliberate attempt means that the attempt is purposefully initiated by the persuader. The
persuader can be a brand, organization or person.
Influence is a change in people’s beliefs, attitudes and behaviors after exposure to
algorithm-mediated communications.
This can vary from:
● Very weak to very strong
● Conscious or unconscious
● With or without consumers’ awareness
● Short term or long term
● Intended or unintended
Online communication is the transmission of algorithm-mediated communication from
senders to receivers online.
The algorithmic persuasion framework is seen as a dynamic process.
5 components of the algorithmic persuasion framework:
● Input
● Algorithm
