A Study on Social Exclusion in Human-Robot Interaction

Round 1

Reviewer 1 Report

I recommend deleting paragraph 0. How to use this template, I think it's a typo of the latex template.

Around line 40 statements are made that are not accompanied by the citation of the sources, I think the sources are [6,7,8,48] in any case we need to add them or in any case improve this aspect. In many points of the work in the first 3 sections there is a lot of talk about existing studies that are not mentioned: in line 57 when it is mentioned that "a robot will better adapt and more accurately predict and match the preferences of the group ...."

 

There is a lot of related work to do

Around line 51 there is talk of trustness but the sources are not cited, in this regard I propose two related articles that could be included in the discussion. [A][B]

 

For better consultation of the text, I recommend numbering the quotations in order of appearance.

In the abstract an experiment in which a bomb is defused is accepted, but this experiment is not mentioned in the introduction.

 

Is John's story in line 137 known (therefore should it be quoted and cited)? Or is it invented. I believe it is invented by the authors but please put it differently because it is not clear if it is a story that must already be known to the background of the reader.

 

Otherwise this article deals with a very hard topic. Social exclusion is a very sad feeling for humans to experience, it is interesting that it has been explored.

 

In tables 1, 2, 3, 4, 5 there are acronyms and abbreviations that may not be known to the reader (for example: M, Me, SD, Sk, etc..) please explain the meaning of these values in the text when we speak of the aforementioned tables.

 

I strongly recommend that you look further into the conclusions. The study carried out has important motivations (presented in the first sections) it is a pity to conclude in a few lines in a hasty and approximate way. Furthermore, please also briefly present in the conclusions the results obtained and the impact that these may have in future applications.

 

[A] doi: 10.1007/978-3-319-47437-3_20

[B] doi: 10.1109/ACCESS.2020.3037701

Author Response

Please see the attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors point out a link between not being understood and being excluded, and being excluded with being more aggressive. The article reports on a study that intends to interrogate these links in the context of an experiment in HRI. This deductive approach, which does not start from data but from hypotheses (clearly formulated) established ex nihilo, comes up against the praxeological reality of human participants engaged in a social activity, as the deployment of the results clearly shows. I think that this article, and especially the results, could turn to your advantage if you were to reflect on your experimental framework itself, i.e. at least on what (i) "teaming up"; and (ii) "talking to a robot as a robot" mean from an interactional point of view. 

Some points in more detail below:

1)

L57 : "So a robot will better adapt and more accurately predict and match the preferences of the group members with a larger data set. This might be perceived as unfairness or even discrimination by other group members. "

>>>Do you mean that humans are, by nature, unfair and discriminant ? For this is exactly how culture, and membership practices work among humans since humanity exists..This what you say actually in the next section, which is very important to remind to the HRI community, in particular the « interdependence issue ».

2)

L186 : "Even though such results provide some clues to how humans behave in discriminating settings while interacting with artificial systems, it is still unclear if they are susceptible to unfair behavior when it is shown by such a system". 

>>I wonder how the results relate to the behavior of a so-called "social" robot as such, or simply that of any technological third party, I mean: how does the fact that it is a robot that distributes resources unfairly to one group rather than another, have a specific effect, because it would be a robot, and not a software for example?

3)

L65 : "In the rest of this paper, we first introduce the topic of humans as social animals by a brief discussion of the evolution of cooperative behavior."

>>I don’t see where you discuss « evolution of ccoperative behavior », the section 2 seems to me to be a review of very basic ideas about membership, but nothing is said about how membership is produced through language and interaction, which is, though, the main resource involved in your very interesting topic on Social Exclusion.I suggest that we dispense with some of the somewhat canonical descriptions of articles, in favor of a somewhat deeper discussion of the practical dimension of exclusion, and the notion of culture.

4)

L252 : "the robot made small talk with one participant" 

>> can you (broadly) characterize what you call « small talk » ?

5)

L275 : "H4. The discriminated participants will feel less as a part of the team in comparison 275 to the favored participants". 

>>are you sure about the word « team » as a a priori relevant characterization ? This not a trivial question since (1) « team » is a practical accomplishment, more than a allegedly given label, and (2) you mention an « experimenter » who is co-present in the room and interact with what you call « the team ». The presence of the experimenter is crucial in the appraisal of the social situation. 

6)

L294, you mention the Woz option to produce your experiment, maybe you should mention it at the very beginning of the paper, and use some space to discuss that (there is a critique literature about WOZ in HRI), because using a won is not anecdotal regarding how you analyze results, and in particular in prospective matters.

7)

L329 : "To make the interaction more realistic, and to help the robot be perceived as a full fledged member of the team, the robot made three suggestions during each game for  changing the ordering of the wires."

>>Could you give arguments that defend the idea that such a conduct of the system (giving a choice) makes the interaction more "realistic"? Because I don't see how that would be the case.

8)

L496 : we arrive at a very interesting section, which should be put forward, and mobilised as a strong methodological argument! The technical problem of address in terms of body orientation is indeed interesting to indicate. Especially since the fact that only 4 groups mentioned it does not mean that it was treated as a problem by the other groups during the experimentation.

L510 : indeed there is a big difference between behavior emerging  in an activity (be it experimental or not) and the talk about this activity later. That is a strong methodological issue.

L520 :" It was as if the robot was the discriminated group member and not the human participants. To explore this phenomenon further requires assessing the attitudes of the participants towards the robot before starting the task. "

This is a very interesting and fundamental issue that deals with categorization processes through interactional conducts, the literature about that is not huge in,  HRI whereas it should be a seminal topic to address ! See for example : 

Alač M (2016) Social robots: Things or agents? AI & Society 31(4): 519-535.

 

Licoppe C and Rollet N (2020) « Je dois y aller ». Analyses de séquences de clôtures entre humains et robot. Réseaux 220-221(2): 151-193.

 

 

9)

L586 : "For all our participants, it was their first encounter with a robot, and they were 586 not sure before the experiment what to expect. "

>>I suggest that it should be mentioned at the very beginning of the paper

 

To me the results you display should lead to a more profound critique of your very experimental protocol, and not just anecdotal aspect of it as it appears in Discussion / conclusion sections. Which could set a brand new methodological perspective for your future works ! 

Author Response

Please see the attached file.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Changes are quite minor, especially regarding the reflexivity towards methodology and  results. But it's better than the previous version.

Author Response

Changes are quite minor, especially regarding the reflexivity towards methodology and  results. But it's better than the previous version.

>> We have added some more text and have added four more references.

Round 3

Reviewer 2 Report

ok