4/29: Iyad Rahwan

[ehuppert]

Comment below with questions or thoughts about the reading for this week's workshop.

Please make your comments by Wednesday 11:59 PM, and upvote at least five of your peers' comments on Thursday prior to the workshop. You need to use 'thumbs-up' for your reactions to count towards 'top comments,' but you can use other emojis on top of the thumbs up.

[kthomas14]

Thank you for sharing your research with us Dr. Rahwan! I was wondering how you believe this research may affect public policy in the future and general sentiment towards autonomous vehicles/vehicles with assisted driving features?

[mikepackard415]

Thank you for sharing your work! Looking forward to hearing more about it tomorrow. I think my classmates have posed a lot of great questions above, so I'll defer to those, but if there is time I would love to hear about your personal academic journey and how you got interested in this topic. Thanks!

[WMhYang]

Thank you very much for sharing your work with us. From my point of view, though understanding the social preferences for morality is crucial, it makes more sense to impose personal preferences in some machines, especially those that are owned only by one person or one principal. For instance, in the self-driving car example, why shouldn't we let the car follow the preferences of the car owner (and I guess machine learning algorithms could figure out personal perferences after learning the characteristics of the owner), which mimics the real world? Thanks.

[YijingZhang-98]

Thanks for sharing with us! I've been thinking about the moral issue concerning AI for a long time, I am so excited to see your excellent work!

[weijiexu-charlie]

Thank you so much for your presentation. Looking forward to your talk tomorrow!

[tianyueniu]

Thank you so much for sharing your work! I look forward to learning more from your talk!

[hihowme]

Thanks a lot for sharing your work! I look forward to hear more about your research! At the same time, I am wondering what do you think would be the key takeaway of your work to other fields like economics and business? Thanks a lot!

[nwrim]

Thanks for coming to our workshop! I remember reading about the moral machine experiment one or two years ago, and I know that a lot of very interesting articles analyzing the result have come out in recent years. I think one of the beauties of this data is that so many people contributed (it is actually large-scale, not "large-scale" in psychology flavor) - do you have any wisdom on how to successfully crowdsourced such large input (so that other interesting social scientific questions can be crowdsourced)?

[YileC928]

Thank you so much for sharing! I look forward to the workshop and hope to see MkramerPsych and SoyBison's questions answered.

[Dxu1]

Thank you for sharing your interesting work! I look forward to hearing your presentation!

[caibengbu]

Thanks for your sharing! Could you elaborate on how did you satisfy the moral standard using this new approach?

[boyafu]

Thanks for sharing your research! Looking forward to the presentation tomorrow.

[timqzhang]

Thank you for your paper ! While I am not quite familiar with this topic, I really look forward to your presentation.

[minminfly68]

Thanks for your sharing. Can you kindly tell us how it could apply to other areas?

[wanxii]

It's a exciting topic. I'm really looking forward to your presentation!

[afchao]

Thank you for presenting to our group! Considering these topics, part of me begins to worry that whatever solutions to ethical dilemmas we end up programming our computers to use (when a decision absolutely must be made) will, over time, erase alternative approaches to those problems such that eventually the nuance of ethics will be forgotten in favor of whatever solution is being applied in practice. Do you share this concern?

[ttsujikawa]

Thank you for sharing your exciting research. It is really interesting to see how you construct ideas where machines possibly fill the gap between ethnic morals and public morals. I look forward to your presentation!!

[luckycindyyx]

Thank you so much for sharing! The moral machines are very interesting topics! My question is similar as @Qiuyu-Li: is the machines only doing the technical work based on the data people feed in, or can they kind of "generate" the new idea of "moral" behavior? Thank you!

[Yutong0828]

Thank you very much for coming and sharing your work with us! I think this is a really interesting topic that studying how machines make decisions, and this reminds me of AIs. Could you please tell us more about how the results from the Moral Machine paper can really help us to design autonomous vehicles? Since for people, moral dilemmas will never have a correct answer, now you have figured out that there are global preferences for moral judgement, then how should we refer to these conclusions when establishing rules for machine behavior?

[97seshu]

Thank you for presenting. Looking forward to learning more about your research.

[jsoll1]

I'm excited for your presentation tomorrow! How do you think we can design more trustworthy machines? And who's qualified to make those decisions? I don't mean like in a government context, but there are many cultures. Should experts on each be consulted for these kinds of questions?

[shenyc16]

Thank you for sharing this interesting research with us. I think figuring out how machines make decisions based on multiple factors is very inspiring. I am still trying to understand whether machines have some humanlike features (e.g. morality). If so, how can we examine it?

[Yiqing-Zh]

Thank you for sharing the work with us. I am looking forward to your presentation.

[XinSu6]

Thank you so much for sharing your work! My question is how do you think the difference in the definition of trustworthy will make a difference in the context of your research?

Looking forward to your presentation.

[ziwnchen]

Thanks a lot for the presentation! The overall framework of machine behaviors & moralities is quite inspiring! My question is, given the fact that machines are rapidly changing, how to make the study of machine behaviors have a long-lasting societal impact?

[chuqingzhao]

Thank you for sharing the work! I am looking forward to your presentation.

[cytwill]

This is really an interesting topic in the field between AI and society. I think I would like to take more time to understand the whole paper. For now, I have some simple questions:

1. Do we have some formal definition of machine morality, and what are the differences between machine morality and human morality?
2. Since we can obtain some moral preference from the experiments, I am wondering how we would use these found patterns to improve machine-based decisions. Are we expecting to make their decisions closer to human decisions? Or are we hoping to debias some moral issues in human's decision system by designing certain algorithms for the machines? And is there any current progress in this perspective?

[siruizhou]

Very excited to have you here! My question is similar to above ones. Look forward to the presentation.

[FranciscoRMendes]

I am very interested to see your presentation on machine morality!

[xxicheng]

Thanks for sharing your work with us!

[j2401]

Hi Dr. Rahwan

I'm curious about how AIs making decisions under uncertainty and randomness. Computer algorithms themselves are lack of randomness and uncertainty, but how should they evaluate and derive information from the random world? Under such conditions, should the accept the outcome that affected by randomness as information, or noise? Thank you for sharing with us and look forward to your presentation!

[qishenfu1]

Thank you for sharing! I am curious that according to your knowledge, what is the biggest advantage of machine learning methods relative to traditional methods? Thank you very much!

[bazirou]

Thank you for sharing the work! I am looking forward to your presentation.

[fyzh-git]

A terrific series of work on the moral machine! Thank you Prof. Rahwan for bringing these enlightening ideas and methodologies to our workshop! In particular, the discussions around the Trolley Problem, its AV version, and the statistical version truly guided the thoughts of readers step by step into a deeper and deeper perspective towards the inherent complex moral issues associated with the machine design. And the opportunities, revealed throughout this process, for the computational social scientists are especially inspiring. I believe they would largely enhance the sense of mission for the future CSS researchers about their efforts on the interdisciplinary work. The design and implementation of the MM platform, which has been performing with a huge success, has fully revealed the advantage of computation-empowered social science work. And I'm only left with two details of concern in terms of the limitations/applicability of the simulated results to be used for informing the real-world scenario: • On MM platform, participants know about the scenario settings, and therefore, being crystal clear about the features of the objects (i.e., age, gender, fitness, social status, number, and species) they are going to assessing, in order to make decisions on the tradeoffs for the Trolley Problem. But in real-world urgent scenario, these features tend to be not so clear by the drivers (or some other decision makers like 'the man in blue' in your second paper). Decision makers tend to prioritize some more obvious features like number and species, rather than the less obvious ones at the first glance like social status. So the semi-informed decisions under restricted time budget might be of large difference from the fully-informed decisions without time limits, and it is very likely to have largely varied re-ordering in the priorities on the features before making the decision. • There might be needs to define a same feature using different metrics under different scenarios, and this could be possible to affect the accuracy and applicability of the simulated results. For example, sometimes variables located within a certain range might be considered homogeneous for a decision maker in urgent scenario; i.e., the driver may not be able to fully consider the accurate number of pedestrian, but tends to know clear about the number of passengers, especially in small cars. (May not be a good example, but I just hope to indicate the concern with metrics choice.) But these are just some of my tiny thoughts arising when reading through the paper. And overall, I personally think your work terrific, from which I do hope to learn a lot! Sincerely look forward to your presentation!

[NikkiTing]

Thank you for sharing your work! I'm also looking forward to the presentation later.

[william-wei-zhu]

Thank you very much for sharing your work with us. We look forward to hearing your presentation.

[wu-yt]

Thank you so much for this interesting research work. Very looking forward to your presentation.

[chentian418]

Hi Dr. Rahwan, Thank you for sharing your research with us! I was curious that, how do you think the sustainability of the Moral machine in predicting the decisions with evolving and varying ethics principles? Look forward to your presentation!

[alevi98]

Dr. Rahwan,

Thanks for coming to our workshop and presenting on one of the biggest questions of our time. Do you think it's a good thing for people to be mistrustful of machines? Insofar as it is, why and how? Insofar as it isn't, how do we improve public opinion of machines?

Best, Alex

[RuoyunTan]

Thanks for sharing your work with us. Looking forward to the presentation!

[ginxzheng]

Thanks for coming Dr.Rahwan!