Computational thinking beyond STEM:an introduction to “moral machines”and programming decision making in Ethics classroom

Antonio Miguel Seoane Pardo

research GRoup in InterAction & eLearning (GRIAL)Research Institute on Educational SciencesDepartment of Education, School Organization and Research MethodsUniversity of Salamanca

aseoane@usal.eshttp://grial.usal.eshttp://twitter.com/aseoane

Technological Ecosystems for Enhancing Multiculturality – TEEM’16University of Salamanca

November 2nd, 2016

Outline

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 2

1. Introduction

2. Computational Thinking and Machine Ethics

3. Training Action

4. Discussion

5. Conclusions and Future Work

1. Introduction

I’m sorry, Dave.I’m afraid I can’t do that

HAL 90002001: A Space Odyssey

Introduction

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 4

• Key competencies at school: different approaches- PISA as a means or as an end?

• The need of fostering STEM skills- More computer science subjects at school?

• STEM skills are not enough- STEM skills are not just a matter of STEM subjects

• We need a holistic approach to face key competencies

Computational Thinking and Machine Ethics

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 5

• “Computational thinking is a way that humans, not computers, think” (Wing, 2006). CT is everywhere

• CT must not be reduced to coding. It has a broader use

• CT should be used as a pedagogical strategy for posing and solving problems, both in sciences and humanities

• We plan to use CT to approach, explain and resolve moral dilemmas of moral machines

Computational Thinking and Machine Ethics

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 6

• Machine Ethics concerns moral behaviour of artificially intelligent artefacts. It is ideal for Ethical discussion

• We intend to apply CT to Ethics itself. Moral approach?- Top-down approach (Kantian deontology, utilitarianism)- Bottom-up approach (Turing, Piaget, Kohlberg)- Hybrid approach

Computational Thinking and Machine EthicsSelf-driving vehicles

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 7

• Experts believe self-driving cars will become commonplace by 2025.

• The main difficulties to overcome are maybe more ethical and sociological than technical.

• What moral approach? Utilitarian, egoistic, profit-based…

• Autonomous vehicles provides us with visual and causal-consequential scenarios for Ethical analysis and discussion

• AV provides us with a testing ground for computational thinking on Ethics

Training actionDecision making model

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 8

• AI programming models (BDI, LIDA) are not feasible(too complicated)

• Other models for the analysis of moral dilemmas do not have “computational structure”

• Our framework: Top-down approach and a finite set of rulesA. Consequentialist approaches

A1. Utilitarianism: “the best action will be that which provides the best or does the least harm”A2. Egoism (self-protection): “the best action will be that which protects me and those who are with me”A3. Profit-based ethics: “the best action will be that which provides the less economic cost”B. Non-consequentialist approaches

B1. Deontology: “the best action will be that which protect those who act according to the rules”B2. Nondeterminism: “behaviour is the result of chance”

Training actionLayout tab

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 9

Title Introduction to moral machines and decision making in Ethics classroom (5 sessions)

Overview

The aim of this activity is to train students of compulsory secondary school in computational thinking applied to Ethics. To do so, the lesson plan focuses on the ethical implications of programming self-driving vehicles to perform actions that will have moral consequences when a crash is unavoidable and damage and harm will be certainly provoked. Students should be able to analyse, represent, study possible outcomes and reflect on ethical behaviours of machines and “program” such behaviours as response to certain inputs, according to ethical principles.

Age 14-16 years Level Medium Skills computational thinking, decision making, logical

reasoning, ethical discussion

Aim of the lesson

This lesson is intended to develop in young people skills related to computational thinking, logic reasoning and algorithmic decision making. This will be done by the analysis and study of different ethical approaches and the formalisation of their main principles to theoretically program machines to perform ethical behaviours. In addition to these “computational” skills, students will become aware of the relevance of Ethics and ethical approaches not only as guiding principles for our daily life decisions, but also to discuss and try to achieve a consensus on some socially accepted moral principles to decide how intelligent agents should behave, both in their interactions with humans, other machines and the environment.

Tools and resources

MIT Media Lab: Moral Machine. http://moralmachine.mit.edu.ClaimMS: AccidentSketch.com. http://draw.accidentsketch.com. AV-DMEC Framework (see lesson plan).Matthieu Cherubini: Ethical autonomous vehicles. https://vimeo.com/85939744.Kahoot: http://kahoot.it.

Training actionSession 1. Introduction and Moral Machine Platform

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 10

Activity 1 Reading and discussing, in groups:“Would you buy a car that should kill you to save other lives?”. El País, June 24th, 2016

Activity 2 How should the car behave? Take your decision and compare with your classmatesSource: http://moralmachine.mit.edu

Training actionSession 2. Analysis of pre-defined scenarios

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 11

Activity 1 Workgroup: Analysis of pre-defined scenarios (as in figure) according to ethical approaches studied in previous lessons

Activity 2 Each group explains their scenario to the class, as so as the ethical approach that better fits into the “better” solution. Is there an agreement in classroom?

Training actionSessions 3-4. Student-designed scenarios

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 12

Activity 1

Workgroup:- Developing their own scenario using the “Matrix with sample elements for developing car crash scenarios” (see table) and applying the Autonomous Vehicle - Decision Making in Ethics Classroom (AV-DMEC) (See figure)- Sketching the scenario with AccidentSketch.com- Describing the scene with logical propositions using connectors

Activity 2 Analyse possible outputs according to the different ethical approaches. What is the best ethical approach according to the “most desirable output”? Why?

Agents AV-car, Bus, School bus,Motorcycle, Cyclist, Pedestrian,Obstacle, Traffic light, […]

Properties Red/green light, Child, Baby,Pregnant woman, Oldman/woman, Wrong way, Sameway, Slower/Faster, with/withouthelmet, Correctly/incorrectlycrossing, Crash, Stop, Run over,[…]

Connectors AND, OR, IF, THEN, ELSEEthicalapproaches

Utilitarianism, Egoism, Profit-based, Deontology,Nondeterministic

Training actionSession5. Discussion and feedback

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 13

Activity 1

Discussion:- Is there any ethical approach that is generally preferable?- Are there scenarios where it should be impossible to determine a “better” output?- Are there scenarios where none of the ethical approaches seem to provide with a reasonable solution? Introduction to complex computational thinking and ethical decision making processes- Matthieu Cherubini: Ethical autonomous vehicles (https://vimeo.com/85939744)

Activity 2 Evaluation:Assessment with game-based learning competition (Kahoot.it!)

Discussion

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 14

• Current models applied to moral machines are yet too complicated to be used in Ethics classroom

• Such models explain how the machine should act, while this experience points on what and why the machine should behave in a certain way

• Current platforms (like MIT’s Moral Machines) analyse user attitudes towards moral machines and there is no a priori ethical approaches

• This proposal let students defining a potentially infinite set of scenarios and analysing them with an ethical framework for decision making

• Students put into practice computational thinking skills, decision making processes, philosophical analysis of moral behaviours, teamwork, oral skills, debating skills, etc.

Conclusions and Future Work

Computational Thinking Beyond STEM: an introduction to “moral machines” and programming decision making in Ethics classroom 15

• Compulsory school should tackle STEM skills from all subjects with cross-curricular approaches

• Computational thinking is not a matter of computer scientists, as ethical decisions are not only for philosophers

• This training action allows students to experience the implications of ethical decisions while putting into practice logical reasoning, resolution of moral dilemmas, etc.

• Future directions of this study will move forward towards the implication of Computer Science subjects: programming scenarios in Scratch, Arduino, etc.

Acknowledgement

Presentation of the TACCLE3 Coding European Project 16

Funded by European Union Erasmus+ KA2 Programme “TACCLE 3 – Coding” (2015-1-BE02-KA201-012307)

This project has been funded with support from theEuropean Commission. This communication reflects the views only of the author, and the Commission cannot be held responsible for any use which may be made of the information contained therein

Computational thinking beyond STEM:an introduction to “moral machines”and programming decision making in Ethics classroom

Antonio Miguel Seoane Pardo

research GRoup in InterAction & eLearning (GRIAL)Research Institute on Educational SciencesDepartment of Education, School Organization and Research MethodsUniversity of Salamanca

aseoane@usal.eshttp://grial.usal.eshttp://twitter.com/aseoane

Technological Ecosystems for Enhancing Multiculturality – TEEM’16University of Salamanca

November 2nd, 2016