Lectures
Lecture 1 What is AI
What is artificial intelligence?
Artificial Intelligence (AI) refers to computer systems that can perform tasks that normally
require human intelligence.
This includes:
● Reasoning (making decisions or solving problems)
● Learning (improving from experience or data)
● Problem-solving (finding solutions to complex tasks)
● Perception (understanding images, speech, or text)
In simple terms: AI tries to make machines “think” or act in intelligent ways similar to
humans.
Key types of AI
● Narrow AI → This is AI designed for one specific task. Examples: image
recognition, spam filters, voice assistants.
This is the type of AI we use today.
● General AI → This would be AI with full human-like intelligence, able to perform
any task a human can do.
This does not exist yet (important for exams!).
● Machine learning → A method where computers learn patterns from data instead
of being explicitly programmed.
Example: a system learns to recognize cats by seeing many cat images.
● Generative AI → AI that can create new content, such as text, images, or code.
Example: tools like ChatGPT.
Why ai matters for society
Economy and work
● AI can automate routine tasks, reducing the need for human labor in some jobs.
● It also creates new jobs that require different skills (e.g. data science).
● AI increases productivity, but can also lead to job loss or inequality.
Key idea: AI changes the job market, both positively and negatively.
Social and democratic life
, ● AI personalizes information (e.g. social media feeds), which can create filter
bubbles and spread misinformation.
● It is used in policing, justice, and welfare, which can affect fairness.
● AI changes how people communicate and form relationships.
Important: AI can influence opinions and democracy.
Health and science
● AI helps discover new medicines faster.
● It improves diagnosis and personalized treatments.
● It supports climate research and scientific discoveries.
AI can strongly improve healthcare and science.
Global power
● AI is a strategic resource, with competition between regions like the US, China, and
EU.
● Not all countries have equal access to AI → leads to global inequality.
● Some countries may become dependent on others for AI technologies.
AI is not just technical, it is also political.
Key Ethical Challenges
Bias & Discrimination
● AI systems trained on historical data can reproduce existing biases.
● This can lead to unfair outcomes in hiring, lending, criminal justice, and healthcare.
AI can reinforce inequality if not carefully managed.
privacy and surveillance
● AI enables large-scale data collection and surveillance technologies like facial
recognition.
● This increases the ability to monitor individuals.
This raises concerns about privacy and civil liberties.
autonomy and manipulation
● Algorithms can influence users’ beliefs and behavior.
● They may exploit psychological vulnerabilities without users being aware.
This can affect individual autonomy and decision-making.
accountability and transparency
● It is often unclear who is responsible for AI decisions.
● Many systems are difficult to understand (“black boxes”).
This makes it hard to challenge or appeal decisions.
,safety and control
● As AI systems become more advanced, they act more autonomously.
● Ensuring they remain safe and under human control is essential.
global inequality
● AI benefits are concentrated in wealthy countries and large corporations.
● This can widen global inequalities.
How can we govern ai?
● Regulation Governments create binding rules. Example: EU AI Act, which restricts
high-risk uses such as social scoring.
● Self-regulation Companies voluntarily adopt ethical guidelines. This is often
criticized for lacking enforcement.
● Technical standards Organizations such as IEEE, ISO, and NIST develop shared
standards.
Code principles of ai governance
● Transparency → AI systems should be explainable
● Accountability → responsibility must be clearly assigned
● Fairness → outcomes should not be discriminatory
● Human oversight → humans should remain in control of important decisions
Ethics of AI main debates (Eliott, ch 8)
● privacy
● manipulation
● opacity (lack of transparency)
● bias
● autonomy
Conclusions
● AI consists of a group of techniques centered around machine learning
● AI is transforming society rapidly
● It is associated with major challenges related to power, including:
○ Bias
○ Inequality
● Effective regulation requires understanding:
○ What AI is
○ How it affects society
○ How regulation interacts with technology, users, and institutions
● Interdisciplinary approaches are essential for addressing AI-related issues
, Lecture 2 AI and journalism
What is AI?
AI (in this context, machine learning) is about finding patterns between input and output.
● The general goal is: Predict an output based on given input features
● During training, the model:
○ Adjusts its internal parameters
○ To minimize prediction errors (make the best possible predictions)
In simple terms: AI learns from examples to make accurate predictions.
ML algorithms: neural networks
Neural networks are inspired by how the human brain works.
● A “neuron”:
○ Receives input values
○ Multiplies them by weights
○ Adds them together → this is called Z
○ Applies a function to produce an output
● Formula: Output = f(β1·x1 + β2·x2 + ...)
● The neuron “activates” (fires) when the total value is high enough.
Important point:
This is mathematically similar to a type of regression model.
What is Ai? Machine learning
Machine learning means:
● Finding the parameters (the “betas”) that minimize prediction errors
● This is similar to minimizing error in statistical models like ordinary least squares
(OLS)
Comparison to traditional statistics:
● Inputs = independent variables
● Output = dependent variable
● Simple neural networks ≈ logistic regression
Key difference:
● Statistics → focuses on understanding relationships (causality)
● Machine learning → focuses on making accurate predictions
Also:
Lecture 1 What is AI
What is artificial intelligence?
Artificial Intelligence (AI) refers to computer systems that can perform tasks that normally
require human intelligence.
This includes:
● Reasoning (making decisions or solving problems)
● Learning (improving from experience or data)
● Problem-solving (finding solutions to complex tasks)
● Perception (understanding images, speech, or text)
In simple terms: AI tries to make machines “think” or act in intelligent ways similar to
humans.
Key types of AI
● Narrow AI → This is AI designed for one specific task. Examples: image
recognition, spam filters, voice assistants.
This is the type of AI we use today.
● General AI → This would be AI with full human-like intelligence, able to perform
any task a human can do.
This does not exist yet (important for exams!).
● Machine learning → A method where computers learn patterns from data instead
of being explicitly programmed.
Example: a system learns to recognize cats by seeing many cat images.
● Generative AI → AI that can create new content, such as text, images, or code.
Example: tools like ChatGPT.
Why ai matters for society
Economy and work
● AI can automate routine tasks, reducing the need for human labor in some jobs.
● It also creates new jobs that require different skills (e.g. data science).
● AI increases productivity, but can also lead to job loss or inequality.
Key idea: AI changes the job market, both positively and negatively.
Social and democratic life
, ● AI personalizes information (e.g. social media feeds), which can create filter
bubbles and spread misinformation.
● It is used in policing, justice, and welfare, which can affect fairness.
● AI changes how people communicate and form relationships.
Important: AI can influence opinions and democracy.
Health and science
● AI helps discover new medicines faster.
● It improves diagnosis and personalized treatments.
● It supports climate research and scientific discoveries.
AI can strongly improve healthcare and science.
Global power
● AI is a strategic resource, with competition between regions like the US, China, and
EU.
● Not all countries have equal access to AI → leads to global inequality.
● Some countries may become dependent on others for AI technologies.
AI is not just technical, it is also political.
Key Ethical Challenges
Bias & Discrimination
● AI systems trained on historical data can reproduce existing biases.
● This can lead to unfair outcomes in hiring, lending, criminal justice, and healthcare.
AI can reinforce inequality if not carefully managed.
privacy and surveillance
● AI enables large-scale data collection and surveillance technologies like facial
recognition.
● This increases the ability to monitor individuals.
This raises concerns about privacy and civil liberties.
autonomy and manipulation
● Algorithms can influence users’ beliefs and behavior.
● They may exploit psychological vulnerabilities without users being aware.
This can affect individual autonomy and decision-making.
accountability and transparency
● It is often unclear who is responsible for AI decisions.
● Many systems are difficult to understand (“black boxes”).
This makes it hard to challenge or appeal decisions.
,safety and control
● As AI systems become more advanced, they act more autonomously.
● Ensuring they remain safe and under human control is essential.
global inequality
● AI benefits are concentrated in wealthy countries and large corporations.
● This can widen global inequalities.
How can we govern ai?
● Regulation Governments create binding rules. Example: EU AI Act, which restricts
high-risk uses such as social scoring.
● Self-regulation Companies voluntarily adopt ethical guidelines. This is often
criticized for lacking enforcement.
● Technical standards Organizations such as IEEE, ISO, and NIST develop shared
standards.
Code principles of ai governance
● Transparency → AI systems should be explainable
● Accountability → responsibility must be clearly assigned
● Fairness → outcomes should not be discriminatory
● Human oversight → humans should remain in control of important decisions
Ethics of AI main debates (Eliott, ch 8)
● privacy
● manipulation
● opacity (lack of transparency)
● bias
● autonomy
Conclusions
● AI consists of a group of techniques centered around machine learning
● AI is transforming society rapidly
● It is associated with major challenges related to power, including:
○ Bias
○ Inequality
● Effective regulation requires understanding:
○ What AI is
○ How it affects society
○ How regulation interacts with technology, users, and institutions
● Interdisciplinary approaches are essential for addressing AI-related issues
, Lecture 2 AI and journalism
What is AI?
AI (in this context, machine learning) is about finding patterns between input and output.
● The general goal is: Predict an output based on given input features
● During training, the model:
○ Adjusts its internal parameters
○ To minimize prediction errors (make the best possible predictions)
In simple terms: AI learns from examples to make accurate predictions.
ML algorithms: neural networks
Neural networks are inspired by how the human brain works.
● A “neuron”:
○ Receives input values
○ Multiplies them by weights
○ Adds them together → this is called Z
○ Applies a function to produce an output
● Formula: Output = f(β1·x1 + β2·x2 + ...)
● The neuron “activates” (fires) when the total value is high enough.
Important point:
This is mathematically similar to a type of regression model.
What is Ai? Machine learning
Machine learning means:
● Finding the parameters (the “betas”) that minimize prediction errors
● This is similar to minimizing error in statistical models like ordinary least squares
(OLS)
Comparison to traditional statistics:
● Inputs = independent variables
● Output = dependent variable
● Simple neural networks ≈ logistic regression
Key difference:
● Statistics → focuses on understanding relationships (causality)
● Machine learning → focuses on making accurate predictions
Also:
