Responsible_CS_DSAIE-01_Introduction.pdf
- 1. ▪ CS4500: Responsible Computer Science DSAIT4500: Responsible Data Science and AI Engineering Lecture 1: Introduction Welcome! 11.02.2025 Anne Fehres and Luke Conroy & AI4Media / Better Images of AI / Data is a Mirror of Us / CC-BY 4.0
- 2. Why Responsible CS/DSAIE? ▪ Assume a non-critical application and the following error: ▪ It doesn't seem like a big issue, right? (as long as it doesn't take too long…)
- 3. Why Responsible CS/DSAIE? ▪ “Move fast and break things” drove digital innovation in the 21th century ▪ Algorithmic systems are increasingly used by governments and companies to make or recommend decisions with significant impacts on individuals, organizations, and society ▪ Real-world impacts: Unfair and discriminatory distribution of services, privacy leaks, impacted autonomy, accidents, unjust crime accusations…
- 5. Responsible CS / Responsible Data Science and AI Engineering ▪ Goal: Develop a responsible attitude that will contribute to future careers. ▪ Organizational, social and technical approaches towards designing and developing Ethical and Responsible DSAIT/CS, ▪ Integrity and ethics in DSAIT/CS and engineering ▪ Interactive and group discussion sessions, based on real-world case studies ▪ Discuss and connect different perspectives related to ethical and responsible DSAIT ▪ 5 EC ▪ Q3
- 6. Course staff Teaching assistants: Marijn Roelvink Arnob Chowdhury Mahmoud M.M.M. Elaref Aleksander Buszydlik Maurya Goyal Mike Segers Popescu Popescu Radu Mihălăchiuţă Teodor Oprescu Marijn Roelvink
- 7. Learning Objectives By the end of this course, you should be able to: LO1: Reflect on your role, responsibilities, and possible dilemmas and trade-offs in developing DSAIT/CS methodologies, both as a professional and as a team member. LO2: Analyse DSAIT/CS approaches and principles towards addressing societal and ethical issues, considering the perspectives of multiple stakeholders. LO3: Contribute to appropriate responsible design, development and impact assessments on DSAIT engineering/ CS work. LO4: Integrate concepts related to trustworthy DSAIT/professional duties and integrity when presenting and justifying your choices to address matters of responsibility connected to DSAIT/CS. From reflecting to integrating and taking ownership.
- 8. Approach ▪ Interactive lectures ▪ Shared lectures, and major-specific lectures ▪ Guest lectures ▪ Reflection on “Responsible content" in courses Q1+2 ▪ Group discussion based on real-world case studies (45 min. after the break) Introduction Integrity Argumentation Moral deliberation Responsible innovation TAI - Explicability TAI - Fairness TAI - Human autonomy TAI - Prevention of harm Entrepreneurial dilemmas Craftship/professionalism Sustainable computing Legal aspects Impact assessment Diversity and Inclusion DSAIT CS DSAIT+CS
- 9. Course Schedule Week Date Time Description Responsible lecturer(s) 1 11-Feb 13:45-15:45 Lecture 1: Introduction Luciano C. Siebert, Cynthia Liem 1 13-Feb 13:45-15:45 Lecture 2: integrity Cynthia Liem 2 18-Feb 13:45-15:45 Lecture 3: Argumentation Matthijs Looij, Cynthia Liem 2 20-Feb 13:45-15:45 Lecture 4: Moral deliberation Cynthia Liem 3 25-Feb 13:45-15:45 Buffer slot 3 27-Feb 13:45-15:45 Lecture 5: Responsible innovation Sietze Kuilman, Luciano C. Siebert 4 4-Mar 13:45-15:45 Lecture 6: Trustworthy AI – Explicability Avishek Anand 4 6-Mar 13:45-15:45 Lecture 7: Trustworthy AI – Fairness (only DSAIT) Cynthia Liem 4 7-Mar 18:00 DEADLINE: Formative assessment 1 - 5 11-Mar 13:45-15:45 Lecture 8: Trustworthy AI – Respect for human autonomy (only DSAIT) Sietze Kuilman, Luciano C. Siebert 5 13-Mar 13:45-15:45 Lecture 9: Trustworthy AI – Prevention of harm (only DSAIT) Sietze Kuilman, Luciano C. Siebert 6 18-Mar 13:45-15:45 Lecture 10: Legal aspects of Responsible DSAIE/CS Guest lecturer, Cynthia Liem 6 20-Mar 13:45-15:45 Lecture 11: Impact assessment Guest lecturer, Cynthia Liem 6 21-Mar 18:00 DEADLINE: Formative assessment 2 - 7 25-Mar 13:45-15:45 Lecture 12: Diversity and Inclusion in DSAIT/CS Sole Pera 7 27-Mar 13:45-15:45 Buffer slot 8 4-Apr 13:45-15:45 DEADLINE: Summative assessment (group report) 10 14 to 18-Apr Group slots to be scheduled Group and individual presentations Team Week Date Time Description Responsible lecturer(s) 1 11-Feb 13:45-15:45 Lecture 1: Introduction Luciano C. Siebert, Cynthia Liem 1 13-Feb 13:45-15:45 Lecture 2: integrity Cynthia Liem 2 18-Feb 13:45-15:45 Lecture 3: Argumentation Matthijs Looij, Cynthia Liem 2 20-Feb 13:45-15:45 Lecture 4: Moral deliberation Cynthia Liem 3 25-Feb 13:45-15:45 Buffer slot 3 27-Feb 13:45-15:45 Lecture 5: Responsible innovation Sietze Kuilman, Luciano C. Siebert 4 4-Mar 13:45-15:45 Lecture 6: Trustworthy AI – Explicability Avishek Anand 4 6-Mar 13:45-15:45 Lecture 7: Entrepreneurial dillemas (only CS) Rini van Solingen 4 7-Mar 18:00 DEADLINE: Formative assessment 1 - 5 11-Mar 13:45-15:45 Lecture 8: Craftship/ Professionalism in CS (only CS) Rini van Solingen 5 13-Mar 13:45-15:45 Lecture 9: Sustainable / Green computing (only CS) Przemek Pawelczak 6 18-Mar 13:45-15:45 Lecture 10: Legal aspects of Responsible DSAIE/CS Guest lecturer, Cynthia Liem 6 20-Mar 13:45-15:45 Lecture 11: Impact assessment Guest lecturer, Cynthia Liem 6 21-Mar 18:00 DEADLINE: Formative assessment 2 - 7 25-Mar 13:45-15:45 Lecture 12: Diversity and Inclusion in DSAIT/CS Sole Pera 7 27-Mar 13:45-15:45 Buffer slot 8 4-Apr 13:45-15:45 DEADLINE: Summative assessment (group report) - 10 14 to 18-Apr Group slots to be scheduled Group and individual presentations Team DSAIT4500 CS4500
- 10. Assessment ▪ Formative (non-graded) - Individual reflection after every group discussion:. Upload to your own Brightspace repository. Support group assignments. Evidence of engagement. - Group work: Self enrolled groups of 4 students. Formative assessment 1: Related to content and discussions around lecture content so far. Feedback. Formative assessment 2: Related to content and discussions around lecture content so far. Peer feedback with other groups.
- 11. Assessment ▪ Summative (graded): PASS/FAIL approach. You will not receive a final grade. Single point rubric system. - Group report (with your existing self-enrolled group) on case study analysis and presentation (P/F) - Individual interview/questions during group presentation (P/F) Both components need to be passed, to receive a pass for the course. Important! Recommended deadline for self-enrolling in a group (4 students!) is February 17 – next week Monday. Step 1) Make use of this and next lecture to talk to people around you: during the breaks and group discussion and try to find a group ☺ Step 2) Use the Brightspace forum to find a group or a missing group member Step 3) Self-enroll your group on Brightspace! Step 4) Meet your group
- 12. Important: note we are running two courses in parallel • DSAIT major follows DSAIT4500, CS major follows CS4500 • This also means we have two (largely parallel) Brightspace courses • Please make sure you enroll in the Brightspace course corresponding to your major… • …and that your self-enrolled group on Brightspace also is in your major’s environment, with fellow students following the same major • You are welcome to also enroll in the parallel course, but we need to register your final assessment under your major’s course code (as that variant is a mandatory core course) • Here in class, you are welcome to mingle with everyone in the room, regardless of major (which even is encouraged!)
- 13. ▪ Answers EWI (https://answers.ewi.tudelft.nl/): • Content-based questions regarding the course • Tags dsait4500 and cs4500 • General guidelines: remain polite, try to be concise, do not share (partial) responses to graded or ungraded assignments, and upvote/accept answers (and questions!) that have helped you ▪ Course email: • For non-content questions specific to you and not relevant to the rest of the class, please contact us via the course email: responsible-dsait-ewi@tudelft.nl and/or responsible-cs- ewi@tudelft.nl ▪ Brightspace: • Problems with Brightspace or other TU Delft tools? Please reach out to Teaching & Learning Support. Course support
- 14. On the in-class group discussions • In responsible computing, it is important to be aware of perspectives that may not be your own, and understand how to navigate these • Note: this does not necessary mean you should always agree! But there should be explicit room for thinking outside of your own box • In this course, we will actively practice this • Within the physical classroom, we will pro-actively encourage you to encounter different perspectives • By creating random groups with those around • By encouraging you to actively think from different perspectives beyond your own
- 15. On the in-class group discussions • But we do understand you may feel more at ease with classmates you know • Hence self-enrollment of a group you choose yourself for the two formative assessment deadlines, and the final summative assessment • Throughout this course, please be empathetic towards one another! • As for discussed cases, some heavy topics may pass by that may trigger strong emotions in some. If so, let’s give each other space for processing these, while at the same time engaging where we can. We purposefully share these with you so our field will hopefully do better in the future. • Please feel safe to take critical stances. At the same time, especially in case of disagreements, please avoid being defensive or offensive – this is a learning setting.
- 16. Example case: personalized content adaptation at Netflix
- 17. Netflix has been personalizing thumbnails • “if you don’t capture a member’s attention within 90 seconds, that member will likely lose interest and move onto another activity” https://about.netflix.com/en/news/the-power-of-a-picture https://netflixtechblog.com/artwork-personalization-c589f074ad76
- 18. What are possible benefits and issues with personalized thumbnails? • Think e.g. of perspectives of: 1. Netflix management 2. A data science researcher at Netflix 3. A customer concerned about tracking and privacy 4. The original producer of the content that is thumbnailed • What would they like? What would they worry about? • How may the technological design take these aspects into account?
- 19. Over the break… • Of course have a break ☺ • Please get a seating ticket for the interactive group session • The seat indicates a zone (A1, A2, A3) and a row number • Row 1 is the front row, row 12 the back row • Upon returning from the break, go to your zone and row • Within the row, sit in pairs with an empty chair in between • Returning your ticket afterwards is highly appreciated row 1 row 2 row 12 …
- 20. Boarding almost completed! • Find the nearest pair of students before/behind you for (row 1, row 2), (row 3, row 4), (row 5, row 6), (row 7, row 8), (row 9, row 10), (row 11, row 12) • This should yield a group of 4 • Introduce yourself and sort your first names in alphabetical order. First name in this order gets a 1, second a 2…etc row 1 row 2 row 12 …
- 21. Group discussion
- 22. What are possible benefits and issues with tailored digital videos for children? • Think e.g. of perspectives of: 1. A parent 2. A babysitter looking after the child 3. An artist producing children’s videos 4. A generative AI enthusiast • What would they like? What would they worry about?
- 23. Different takes on children’s content https://www.youtube.com/watch?v=020g-0hhCAU https://www.newyorker.com/magazine/2024/06/17/cocomelon-children-television-youtube-netflix https://www.theintrinsicperspective.com/p/here-lies-the-internet-murdered-by
- 24. Different takes on children’s content https://www.youtube.com/watch?v=6dWXexP_WMk
- 25. If you were in charge of an online video platform, what would you consider a responsible take on personalizing and recommending children’s videos? • First think of this yourself • Then discuss in the group
- 26. For your individual Brightspace submission • Write a short reflection (2 pages max) on this session and your expectations for the course: • Anything you particularly wish to learn about in the course? • How easy/challenging was it for you and your group to consider the given cases from the given perspectives? • Did you have enough time and space to have a proper discussion? • What is your take on the personalized thumbnail in Netflix, and tailored children’s content creation?
- 27. ▪ CS4500: Responsible Computer Science DSAIT4500: Responsible Data Science and AI Engineering Lecture 1: Introduction Welcome! 11.02.2025 Anne Fehres and Luke Conroy & AI4Media / Better Images of AI / Data is a Mirror of Us / CC-BY 4.0 Thanks for returning your seat tickets!