Ai and human decision-making
A.A. 2024/2025
Obiettivi formativi
Aim of the course are: a) to build a foundational knowledge of the main features of human learning, reasoning and decision making, their unique strengths and weaknesses compared to current AI, and how AI can support (vs endanger) them; b) the legal provisions, principles, and concepts that may shield individuals from the most severe dangers of the digital environment.
Risultati apprendimento attesi
Unit 1
Knowledge and understanding
· Understand the generalities of human cognitive processes as interdependent functions evolved and adapted to a natural physical and social environment
· The multi-layered human intelligence: automatic vs controlled processes (both for learning and reasoning)
· What's hard for humans and easy for machines, what's easy for humans and hard for machines
Applying knowledge and understanding
· How AI can integrate and support human judgments
· How AI can mislead and endanger human judgments
Unit 2
Knowledge and understanding
· Understand the ideal standards of decision-making both in the individual and interactive context
· Understand why people fail to cope with ideal standards
· Heuristics in decision-making and associated biases
· Prospect theory and associated formal modeling of decision-making
· Understand how indirect suggestions can influence decisions (nudging)
· Human metacognitive abilities and their limits
· Optimal advice integration and human departures from optimality
Applying knowledge and understanding
· Determination of the optimal course of action in different contexts, with examples from clinical decision-making and economic decisions
· Analysis of the typical decision course of individuals, with critical analysis of their limits
· Analysis of human advice integration, with critical analysis of their limits
Unit 3
Knowledge and understanding
· Contextualized knowledge of foundational legal principles and concepts related to data law and technology regulation
· Insight into the challenges that digitization poses for the legal environment
· Knowledge of possible policy solutions to the future of data regulation and new technologies
Applying knowledge and understanding
· The ability to critically evaluate key concepts of data laws, including the EU General Data Protection Regulation (GDPR), the IP, the Data Act, and the Artificial Intelligence Act
Knowledge and understanding
· Understand the generalities of human cognitive processes as interdependent functions evolved and adapted to a natural physical and social environment
· The multi-layered human intelligence: automatic vs controlled processes (both for learning and reasoning)
· What's hard for humans and easy for machines, what's easy for humans and hard for machines
Applying knowledge and understanding
· How AI can integrate and support human judgments
· How AI can mislead and endanger human judgments
Unit 2
Knowledge and understanding
· Understand the ideal standards of decision-making both in the individual and interactive context
· Understand why people fail to cope with ideal standards
· Heuristics in decision-making and associated biases
· Prospect theory and associated formal modeling of decision-making
· Understand how indirect suggestions can influence decisions (nudging)
· Human metacognitive abilities and their limits
· Optimal advice integration and human departures from optimality
Applying knowledge and understanding
· Determination of the optimal course of action in different contexts, with examples from clinical decision-making and economic decisions
· Analysis of the typical decision course of individuals, with critical analysis of their limits
· Analysis of human advice integration, with critical analysis of their limits
Unit 3
Knowledge and understanding
· Contextualized knowledge of foundational legal principles and concepts related to data law and technology regulation
· Insight into the challenges that digitization poses for the legal environment
· Knowledge of possible policy solutions to the future of data regulation and new technologies
Applying knowledge and understanding
· The ability to critically evaluate key concepts of data laws, including the EU General Data Protection Regulation (GDPR), the IP, the Data Act, and the Artificial Intelligence Act
Periodo: annuale
Modalità di valutazione: Esame
Giudizio di valutazione: voto verbalizzato in trentesimi
Corso singolo
Questo insegnamento può essere seguito come corso singolo.
Programma e organizzazione didattica
Edizione unica
Periodo
annuale
Programma
● Unit 1.
○ What is "Human Centered AI"
○ Human thinking and reasoning:
■ Associative learning and thinking (system 1)
■ Analytical reasoning
■ At the junction between associative and analytical thinking: causal reasoning
○ Current AI:
■ Transparent/opaque algorithms
■ Associative AI vs model-based AI
■ Explainability
○ How AI can help human thinking, and how humans can exploit other humas by using AI
● Unit 2.
1. Understanding Rules
- Different Types of Rules
- The Dynamics of Rule Application
- The Essentials of Required Rules
2. The Nature and Function of Rules
- The Role of Constitutive and Regulatory Rules
- The Rationality Behind Norms
- Legal Principles vs. Rules
- Challenges in Rule Terminology
3. Language and Rules
- The Boundaries of Common Language in Rule Contexts
4. Generalization in Rule Application
- General Rules and Their Prescriptive Nature
5. Theoretical Underpinnings of Rules
- The Conceptual Basis for Adhering to Rules
- The Craft of Rule Development and Standardization
6. The Structure of Rule Systems
- The Stratification of Rule Layers
7. Rules as Guides
- How Rules Serve as Justifications
8. The Influence of Rules
- The Origin and Power of Normative Influence
- How Rules Gain Significance and Impact
- The Uneven Distribution of Rule Authority
● Unit. 3.
○ Single agent human decision making
■ Choice under certainty
■ Judgment under risk and uncertainty
■ Choice under risk and uncertainty
■ Prospect theory, nudging and boosting
■ Perceptual decision making
○ Human multi-agent decision making
■ Strategic interaction
■ Social decision making
○ Human - AI collaborative decision making
○ What is "Human Centered AI"
○ Human thinking and reasoning:
■ Associative learning and thinking (system 1)
■ Analytical reasoning
■ At the junction between associative and analytical thinking: causal reasoning
○ Current AI:
■ Transparent/opaque algorithms
■ Associative AI vs model-based AI
■ Explainability
○ How AI can help human thinking, and how humans can exploit other humas by using AI
● Unit 2.
1. Understanding Rules
- Different Types of Rules
- The Dynamics of Rule Application
- The Essentials of Required Rules
2. The Nature and Function of Rules
- The Role of Constitutive and Regulatory Rules
- The Rationality Behind Norms
- Legal Principles vs. Rules
- Challenges in Rule Terminology
3. Language and Rules
- The Boundaries of Common Language in Rule Contexts
4. Generalization in Rule Application
- General Rules and Their Prescriptive Nature
5. Theoretical Underpinnings of Rules
- The Conceptual Basis for Adhering to Rules
- The Craft of Rule Development and Standardization
6. The Structure of Rule Systems
- The Stratification of Rule Layers
7. Rules as Guides
- How Rules Serve as Justifications
8. The Influence of Rules
- The Origin and Power of Normative Influence
- How Rules Gain Significance and Impact
- The Uneven Distribution of Rule Authority
● Unit. 3.
○ Single agent human decision making
■ Choice under certainty
■ Judgment under risk and uncertainty
■ Choice under risk and uncertainty
■ Prospect theory, nudging and boosting
■ Perceptual decision making
○ Human multi-agent decision making
■ Strategic interaction
■ Social decision making
○ Human - AI collaborative decision making
Prerequisiti
Unit 1: since we'll start with "high level" human cognitive processes (thinking and reasoning), students should have foundational knowledge of basic cognitive processes (mainly perception, attention and memory). Those who do not have a bachelor in psychology should study a good textbook on cognitive processes, limited to those chapters. A good basic textbook in English is Sternberg's "Cognitive psychology", 7th edition. A basic, high-school level understanding of probability calculus and propositional logic is also a necessary requirement (use whatever textbooks you have available).
Metodi didattici
Unit 1. Flipped classroom & Problem-based learning. Students are required to study the materials discussed each week before the lectures. The lectures will unfold as discussions, explanations, and problem-solving tasks on some of the issues illustrated in the textbooks.
Unit 2. Flipped classroom & Problem-based learning. Students are required to study the materials discussed each week before the lectures. The lectures will unfold as discussions, explanations, and problem-solving tasks on some of the issues illustrated in the textbooks.
Unit 3. Lectures, short movies, classroom discussions, group work, and exercises. Smartphone apps that allow students to respond in real-time to open or closed questions will be used.
Unit 2. Flipped classroom & Problem-based learning. Students are required to study the materials discussed each week before the lectures. The lectures will unfold as discussions, explanations, and problem-solving tasks on some of the issues illustrated in the textbooks.
Unit 3. Lectures, short movies, classroom discussions, group work, and exercises. Smartphone apps that allow students to respond in real-time to open or closed questions will be used.
Materiale di riferimento
Unit 1. Chosen chapters from (consult the course website one month before course starts for knowing which ones to study for the first week):
Pearl, J, MacKenzie, D. (2018). The Book of Why: The New Science of Cause and Effect. Basic Books.
Khaneman, D. (2011). Thinking, fast and slow. Farrar, Strauss, and Giroux.
Johnson-Laird, P (2009). How we reason. Oxford University Press.
Chetouani, M., Dignum, V., Lukowicz, P., Sierra, C. (eds) (2023). Human-centered Artificial Intelligence: advanced lectures. Springer.
Whole books (reading materials, more than studying materials):
Broussard, M. (2018). Artificial unintelligence: how computers misunderstand the world. MIT press.
Alter, A. (2017). Irresistible: the rise of addictive technology and the business of keeping us hooked. Penguin press.
Unit 2. The primary text of reference, but not the only one, will be:
Frederick F. Schauer, Playing by the Rules: A Philosophical Examination of Rule-Based Decision-Making in Law and in Life (1991). To prepare for the exam, students will also need to be familiar with everything that will be published on the e-learning site during the course.
Unit 3. Chosen chapters from (consult the course website one month before course starts for knowing which ones to study for the first week):
Angner, E. (2020). A Course in Behavioral Economics (Third edition.). London: Palgrave.
Hunink, M. G. M. (2014). Decision Making in Health and Medicine: Integrating Evidence and Values (2° edizione). Cambridge University Press.
Khaneman, D. (2011). Thinking, fast and slow. Farrar, Strauss, and Giroux.
Further compulsory material will be made available by the teacher during the course.
Pearl, J, MacKenzie, D. (2018). The Book of Why: The New Science of Cause and Effect. Basic Books.
Khaneman, D. (2011). Thinking, fast and slow. Farrar, Strauss, and Giroux.
Johnson-Laird, P (2009). How we reason. Oxford University Press.
Chetouani, M., Dignum, V., Lukowicz, P., Sierra, C. (eds) (2023). Human-centered Artificial Intelligence: advanced lectures. Springer.
Whole books (reading materials, more than studying materials):
Broussard, M. (2018). Artificial unintelligence: how computers misunderstand the world. MIT press.
Alter, A. (2017). Irresistible: the rise of addictive technology and the business of keeping us hooked. Penguin press.
Unit 2. The primary text of reference, but not the only one, will be:
Frederick F. Schauer, Playing by the Rules: A Philosophical Examination of Rule-Based Decision-Making in Law and in Life (1991). To prepare for the exam, students will also need to be familiar with everything that will be published on the e-learning site during the course.
Unit 3. Chosen chapters from (consult the course website one month before course starts for knowing which ones to study for the first week):
Angner, E. (2020). A Course in Behavioral Economics (Third edition.). London: Palgrave.
Hunink, M. G. M. (2014). Decision Making in Health and Medicine: Integrating Evidence and Values (2° edizione). Cambridge University Press.
Khaneman, D. (2011). Thinking, fast and slow. Farrar, Strauss, and Giroux.
Further compulsory material will be made available by the teacher during the course.
Modalità di verifica dell’apprendimento e criteri di valutazione
Unit 1. Written exam: open and closed choice questions. Students are strongly encouraged to take the assessment immediately at the end of the module.
Unit 2. Written exam: open and closed choice questions. Students are strongly encouraged to take the assessment immediately at the end of the module.
Unit 3. Written exam: open and closed choice questions. Students are strongly encouraged to take the assessment immediately at the end of the module.
Unit 2. Written exam: open and closed choice questions. Students are strongly encouraged to take the assessment immediately at the end of the module.
Unit 3. Written exam: open and closed choice questions. Students are strongly encouraged to take the assessment immediately at the end of the module.
M-PSI/01 - PSICOLOGIA GENERALE - CFU: 12
Lezioni: 96 ore
Docenti:
Cherubini Paolo, Reverberi Franco, Rossetti Andrea
Siti didattici