5/6: Michal Kosinski

[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.

[YileC928]

Thanks for sharing Dr. Kosinski, I am also quite interested in hearing your thoughts on HireVue. Looking forward to the workshop.

[luckycindyyx]

Thanks for sharing such interesting paper with us! My questions are: what are the techniques that can lower the risk of facial recognition technologies? And are the law and regulation mature enough to punish any relevant risk? Thank you!

[FrederickZhengHe]

Thanks for sharing such an interesting paper! I am perplexed by many points, so I look forward to listening to your presentation tomorrow!

[Qlei23]

Hi Dr. Kosinski, thank you for sharing your work with us and for coming to present at our workshop! I really enjoy reading your paper and I also know your research from the famous "Gaydar" study. I want to hear your opinions on the ethical issues of these findings and on privacy concerns of the widely used facial recognition technology. Thank you and looking forward to your presentation!

[weijiexu-charlie]

Thanks for your presentation! Looking forward to your talk tomorrow.

[ghost]

Could you please explain psychological profiling in more detail?

[zixu12]

Thanks for sharing your work with us! My question is also related to moral concern. In particular, how do you collect these image data? Is there any procedure that or do you have any suggestions on what we had better follow?

[jsoll1]

Can you be more specific on the ethnic distribution of your participants? You mentioned that like a fifth of your sample is non-white, but I'm curious as to which ethnicities are represented in that.

[chun-hu]

This is an interesting study Dr. Kosinski! I look forward to your presentation.

[Leahjl]

Thanks for presenting at our workshop!

[shenyc16]

Thank you for sharing this interesting work with us. My question is on moral hazards or some potential threats to ethics of facial recognition. I wonder wthether more and more specific research in this field may further exacerbate social tensions on ethnic, race or gender. If so, do we have a proven solution?

[Yutong0828]

Hi Professor Kosinski, it's a great surprise and pleasure to have you here! I have been really interested in using digital data to predict one's characteristics, especially personality traits. However, one concern is that the noise in the data and the representativeness of the sample. Since in psychology, we are normally able to ensure similar experiment environment for each participant, but in big data study, we cannot control that. I understand that by enlarging the data set, we shall be able to obtain more robust result, but as a student, I always find it hard to obtain enough amount of data. So, I was wondering for psychological study, how we may obtain enough data and reduce the impact of noisy data. Thank you very much!

[XinSu6]

Thank you so much for sharing your work!

I wonder since different people and regions would treat ethics differently, would that be a problem for the purpose of your research?

Thank you and Looking forward to the presentation.

[RuoyunTan]

Thank you for sharing your work with us. Looking forward to the presentation tomorrow.

[NaiyuJ]

Thanks for this super interesting work! I'm also curious about the ethical concern of the further implication of the results. Look forward to your presentation!

[luyingjiang]

Thank you for sharing your work. I am curious about the privacy issue when collecting people's information. Looking forward to it.

[wanxii]

Thanks for sharing your work and really look forward to your presentation!

[romanticmonkey]

Thank you so much for your presentation! I wonder how will you prevent such method/technology from falling into the wrong hands? The implication of this study rings the bell of many dystopia literature and TV shows.

[heathercchen]

Thank you for your presentation, Prof. Kosinski! I am wondering about your opinions towards privacy in digital information collection, especially when it comes to people's behaviors, such as facial expressions, movements, digital traces and so on. Do you think requiring such information more and more will inevitably lead to a situation that invades more into people's private zones? Thanks a lot!

[wanitchayap]

I would be extremely happy if you answer Regina questions! She really nails them!! Thank you :)

[NikkiTing]

Thank you for sharing your work! As already mentioned by many others, I am also interested in the ethics and privacy concerns with respect to your research. I look forward to your presentation.

[nwrim]

I am looking forward to the discussion on ethical concerns that my peers raised (greatly) on previous questions. In addition, I am wondering if there is some intuition on what aspect the black box models actually capture - I know that prediction over explanation is one important trend of social science researches nowadays, but it will be great if we can see some explanation, especially in regard with features in these types of research.

[caibengbu]

Thank you for sharing your work with us, Dr. Kosinski. My question is how your work could be applied to other fields such as economics.

[YijingZhang-98]

I really enjoy reading your work! Look forward to your presentation! Thanks!

[xzmerry]

Thank you for sharing! The research on leveraging facial recognition technology on political orientation detection sounds interesting, but as many people have mentioned, I am also curious about its possible ethical issues.

Besides, as your research also covers digital traits related to individuals' behavior, I am looking forward to your presentation and to see how we could incorporate this into researches about behavioral economics.

[boyafu]

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

[97seshu]

Thank you for presenting. Hope to learn more about your work tomorrow.

[hihowme]

Thanks a lot for sharing your work! I look forward to your presentation tomorrow!

[chentian418]

Thank you for presenting at our workshop, Dr Kosinski! I am interested in your study in comparison of personality judgement between computer models and human.

Firstly, would you mind extending in terms of psychology and computational social science that how participants' likes on Facebook link to their personality and traits, and I'm interested in more evidence of making psychological inference based on AI and big data.

Moreover, I'm curious that if there's any problem about taking self-rating of participants as dependent variables, as it seems that there may be some inherent bias in people's self-evaluation in their personality and their likes on Facebook, and personality judgement from their friends may not necessary deviate a lot from their true personality even if the prediction accuracy using self-rating as explanatory variables shows so.

Look forward to seeing you tomorrow!

[siruizhou]

Thank you for sharing. I am interested in learning more about the bias that can be introduced through categorization.

[fyzh-git]

Thank you Dr. Kosinski for presenting at our workshop. One method that is really smart is to move from the absolute classification accuracy to the accuracy offered by alternative ways. It is very useful and effective in circumventing the inherent reference group effect under the self-reported labels!

But I would challenge the reliability of the comparison established in the second paper, Computer-based personality judgments are more accurate than those made by humans. The choice of Facebook friends as the proxy of human personality judgments might itself contains a large bias, especially when only one or two are randomly chosen for generating the results.

I think such a choice could only be valid under the assumption of ''closed' friends, but how could this be guaranteed? Thus, the outcome that algorithms outperform human predictions based on the above 'human comparison' might not be so persuasive for the predictability of algorithms. Would you mind explain a little bit of the sample selection process? Thank you!

[ziwnchen]

Thanks for the presentation! I am curious about the general implication of using big data and computational methods to better understand personality traits. It's interesting to see that computer models could more accurately predict personalities than human judgment. However, the "personality" is still measured using pre-established categories. Do you think big data might have the potential to better inform us of a more fundamental structure of human personality or other psychological mechanisms?

[Yiqing-Zh]

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

[WMhYang]

Thank you very much for sharing your work. In the facial recognition paper, we see that facial structures have strong prediction power of the political orientation of hte individuals. I was wondering what mechanism drives this result, i.e. which kinds of facial structure are associated with which sets of psychological factors or environmental factors, so these factors drive the political orientation. Thanks again.

[timqzhang]

Thank you for your work! It is really exciting and look forward to your presentation !

[j2401]

Thank you Dr. Kosinksi for sharing your work with us. l'm also looking forward you sharing insights about the dilemma of AI technology and privacy issues!

[cytwill]

Thank you for presenting this interesting work. I am interested in the ethical issues of doing digital psychological experiments and some advancements of AI techniques in psychometric studies.

[Dxu1]

Thank you for presenting this interesting work. I am interested in your thought as to what would be the most practical application of such facial algorithm for detecting political orientation.

[yierrr]

Thanks for presenting! I'm also curious about the potential legislation problems about the use of facial recognition technologies. Thank you!

[ttsujikawa]

Thank you for your presentation! I am curious about how you treat regional differences in your analysis since people in different regions have significant variations in their notions of ethics.

Thank you!

[Bin-ary-Li]

Hi, sorry if I being very blunt here but I want to address the elephant in the room, i.e., the controversial paper on deep learning prediction of homosexuality. Do you think the original conclusion still hold? Would you agree that the publication of the paper took advantage of its reviewers who are in the field of social psychology and might not be proficient in the methodological assumption of ML?

[chuqingzhao]

Thank you Dr. Kosinksi for sharing your work. I look forward to the presentation.

[bazirou]

Thank you for your work! It is really exciting and look forward to your presentation !

[bowen-w-zheng]

Thanks for presenting. Look forward to how you address the controversy in your works.

[hhx2207061197]

Thank you Dr. Kosinksi for sharing your work. I look forward to the presentation.

[qishenfu1]

Hi Prof. Kosinski, thank you for sharing your wonderful work! I am just curious that what do you think about the privacy issue of this topic?

[ddlee19]

Looking forward to the presentation!

[william-wei-zhu]

Hi Professor Kosinski, looking forward to your presentation soon!