=Paper=
{{Paper
|id=Vol-3276/SSS-22_FinalPaper_15
|storemode=property
|title=Should Social Robots in Retail Manipulate Customers?
|pdfUrl=https://ceur-ws.org/Vol-3276/SSS-22_FinalPaper_15.pdf
|volume=Vol-3276
|authors=Oliver Bendel,Liliana Margarida Dos Santos Alves
|dblpUrl=https://dblp.org/rec/conf/aaaiss/BendelA22
}}
==Should Social Robots in Retail Manipulate Customers?==
https://ceur-ws.org/Vol-3276/SSS-22_FinalPaper_15.pdf
Should Social Robots in Retail Manipulate Customers?
Oliver Bendel and Liliana Margarida Dos Santos Alves
School of Business FHNW, Bahnhofstrasse 6, CH-5210 Windisch
oliver.bendel@fhnw.ch; alves.liliana1989@gmail.com
Abstract Switzerland has been a pioneer in this field, with several
Against the backdrop of structural changes in the retail trade, projects in Zurich and the surrounding area. Other locations
social robots have found their way into retail stores and shop- involved in this include California, Japan, and Germany
ping malls in order to attract, welcome, and greet customers; where people rely on relevant social robots (Bendel 2021b).
to inform them, advise them, and persuade them to make a
Consultants and salespeople often have broad knowledge
purchase. Salespeople often have a broad knowledge of their
product and rely on offering competent and honest advice, of their product and rely on competent and, all in all, honest
whether it be on shoes, clothing, or kitchen appliances. How- advice, whether it be on shoes, clothing, toys, or tools. How-
ever, some frequently use sales tricks to secure purchases. ever, some of them frequently use sales tricks and attempt
The question arises of how consulting and sales robots should to outsmart customers to secure purchases. They persuade
“behave”. Should they behave like human advisors and sales-
the customer, for instance, that a shirt or blouse looks good
people, i.e., occasionally manipulate customers? Or should
they be more honest and reliable than us? This article tries to on him or her, or that he or she urgently needs a certain
answer these questions. After explaining the basics, it evalu- kitchen appliance, which may not be the truth in reality. To
ates a study in this context and gives recommendations for boost sales, they use certain tricks and strategies they have
companies that want to use consulting and sales robots. Ulti- learned in specialized sales training courses. Over the course
mately, fair, honest, and trustworthy robots in retail are a win-
of their professional life, additional skills are added.
win situation for all concerned.
Several questions arise from this: How should consulting
and sales robots “behave”? Should they behave like human
Introduction advisors and salespeople by occasionally manipulating cus-
tomers? Or should they instead be more honest and reliable
A structural change is reshaping the retail trade. Stationary than humans? What do customers really want? Could it be
trade is losing ground to e-commerce. There are attempts to that social robots are conceded more than people when they
regain competitive parity and reverse this development with utter half-truths or deceptions? After all, they have no con-
the use of online retailing through onsite technologies (in- science – which in turn raises the question of the conscience
cluding sales and consulting or advisor robots). Robots are of the (actual responsible) people behind the sales robot.
seen as having great, but as yet untapped, potential. For ex- While there are many studies on how social robots should
ample, it would be possible to operate 24/7 and make cost behave in principle (Bendel 2021a), and some articles on so-
savings by substituting personnel. A higher productivity and cial robots in retail (Aaltonen et al. 2017), how social robots
profit maximization of the stationary trade is conceivable, should behave specifically in retail has been rarely exam-
furthermore better advisory services, since more extensive ined. This paper explores the question of whether social ro-
and more current information can be called up by robots bots in retail should manipulate customers or be fair, honest,
than people. Years ago, social robots such as Pepper, NAO, and trustworthy. First, it outlines how social robots are used
and Cruzr made their debut to welcome, inform, advise cus- in retail stores today. Then, their AI-related capabilities are
tomers, and ultimately persuade them to make a purchase. presented. The next section summarizes a study conducted
Social and, particularly, humanoid robots often function by Liliana Alves as part of her master’s thesis. She surveyed
as service robots, but go far beyond classical models (think over 300 people on whether consulting and sales robots
of cleaning robots for the floor or for windows), in their de- should manipulate customers. Finally, recommendations are
sign, their natural language capabilities, and other functions given for companies on how to use social robots in retail.
of artificial intelligence (AI) like face and voice recognition. The articles rounds off with a summary and outlook.
___________________________________
In T. Kido, K. Takadama (Eds.), Proceedings of the AAAI 2022 Spring Symposium
“How Fair is Fair? Achieving Wellbeing AI”, Stanford University, Palo Alto, California,
USA, March 21–23, 2022. Copyright © 2022 for this paper by its authors. Use permitted
under Creative Commons License Attribution 4.0 International (CC BY 4.0).
8
Social Robots in Retail 2017). They welcome customers, who can approach them
and ask for information. If one of the models is over-
The following section discusses social robots in retail. First, whelmed by this interaction, an employee is switched on via
the terms “social robot” and “service robot” are clarified. its display on the chest to provide the requested information.
Then, examples of robots in retail are given. Finally, their At MediaMarkt and Saturn, several models were trialled,
functions are discussed, especially AI-based capabilities. such as Pepper and NAO and particularly, Paul (Dinske
2018). This is basically a Care-O-bot from the Fraunhofer
Social Robots and Service Robots Institute for Manufacturing Engineering and Automation
Social robots are sensorimotor machines created to interact IPA in Germany, originally intended for nursing and care,
with humans or animals, particularly more sophisticated except the arms were removed. Paul was active in the
species (Bendel 2021a). They can be determined through Sihlcity shopping mall in Zurich, among other places.
five key aspects. These are: interaction with living beings, Early on, California and Japan also experimented with so-
communication with living beings, proximity to living be- cial robots in retail. As of 2016, a Pepper was available to
ings, representation of (aspects of or features of) living be- answer questions at the Westfield San Francisco Centre
ings (e.g., they have an animaloid or a humanoid design or (Evangelista 2016). In Japan, it also popped up in several
natural language abilities), and fundamentally, utility for stores shortly after its “birth”. The reports on this repeatedly
living beings. A broad definition covers software robots as sounded alarmist: “Tokyo firm replaces staff with a team of
well as hardware robots, and so could include certain chat- Pepper the ‘emotional’ humanoids.” (Woollaston 2016) – so
bots and voice assistants, relativizing the sensorimotor as- claimed Mail Online on March 24, 2016.
pect. Social relationships are often one-to-many relation-
ships, not just one-to-one. In retail, both occur – but the so-
cial robot can usually only address one person for technical
reasons. This can make the others present, such as the cus-
tomer’s partner or friend, feel uncomfortable.
Some social robots are service robots, that is, they handle
certain services and provide certain assistance, and con-
versely, some service robots are also social robots, insofar
as they have communication and interaction functions. Typ-
ical examples in this intersection are care and therapy ro-
bots, but also advisor and sales robots in the retail trade. In
this paper, they are seen as social robots that offer certain
service functions such as informing and advising a customer
or processing their purchase (Meyer et al. 2020a).
Social robotics crosses over with machine ethics. Moral
and immoral conversational agents have emerged from this
discipline in recent years – including a chatbot that can sys-
tematically lie (Bendel et al. 2017).
Examples for Robots in Retail
Robots of all kinds are appearing in the retail industry (Kelly
2020; Meyer et al. 2020a/b). Some are transport robots like
Relay from Savioke, which moves goods through the aisles
of a hardware store, others are security robots like K5 from
Knightscope, which patrols the grounds of companies and Fig. 1: K5 (Stanford Shopping Center, 2016)
alliances (see Fig. 1). Tory was “employed” as an inventory
assistant at Adler Modemärkte AG in Germany (Bendel A kangaroo-like robot has an unusual use in Tokyo. “In
2021b). Social robots serve as caregivers and as toys, for August, a robot vaguely resembling a kangaroo will begin
example in shopping mall nurseries. Some also function as stacking sandwiches, drinks and ready meals on shelves at a
consulting and sales robots. Only in a few cases are they re- Japanese convenience store in a test its maker, Telexistence,
sponsible for the tasks mentioned above, such as transporta- hopes will help trigger a wave of retail automation.” (Kelly
tion and security. Selected examples are given below. 2020) Whether it is to be understood as a social robot in the
Several Pepper robots have been deployed in Zurich’s narrower sense would have to be examined more closely.
Glatt shopping mall since 2017 (Vontobel and Weinmann Without any doubt, it is a service robot.
9
Functions of Social Robots in Retail
The social robot in retail uses its appearance to attract the
customer’s attention and to entertain him or her. It often has
a humanoid design, at least in the form of a larger body and
two arms. Often they will have eyes and a mouth that change
color or move. The face is particularly important for cus-
tomer engagement (Song and Luximon 2021). Forms of ro-
bot enhancement are possible, such as donning a uniform or
wig.
Body and arm movements of the social robot appear en-
tertaining and can encourage imitation – think of the danc-
ing movements of NAO or the therapeutic movements of P-
Rob or Lio by F&P Robotics in Switzerland (Bendel 2018).
In addition, they serve as nonverbal communication. Cus-
tomers appreciate being able to do the fist bump or high five Fig. 2: Pepper’s Hand (Westfield, 2017)
with Pepper. Prasad et al. (2021) refer to this as human-robot
handshaking. Its hand can hardly grasp anything – but is de- Another function is AI-supported analysis during the fit-
signed very naturally so that the greeting also appears natu- ting of clothes. The system can, for instance, assess the fit
ral (see Fig. 2). of the garments and the matching of colors. It can also assist
Natural language capabilities are important to enable in- with a virtual fitting – so far, this has mainly been important
tuitive operation and to speak to the customer as a human in online retail, i.e. in situations where the garments are not
consultant or salesperson would. The voice is crucial here, physically available for a fitting (Werdayani and Widiaty
and it should sound pleasant and convincing. This is a chal- 2021). In this context, the virtual fitting and analysis of the
lenge for robots like Pepper, which inherently have a ro- physical fitting could be combined.
botic, childlike voice. Speech models such as GPT-2 and In principle, video presentations, text, and image infor-
GPT-3, used in social robots such as Harmony, allow for mation are transmitted via an integrated display or via natu-
longer conversations (Coursey 2020). Whichever speech ral language. This can be classically acquired data, but also
model is used, they must be furnished with company-spe- data acquired through machine learning and deep learning
cific coordinates, and a knowledge base with appropriate capabilities. Thus, in conjunction with appropriate AI sys-
data about the location and availability of products. tems, the social robot can learn from conversations with cus-
Some models used in retail are capable of face and voice tomers and from their behavior and apply this to new con-
recognition as well as gesture recognition, and some are ca- tacts. If it is connected to other systems, such as the booking
pable of emotion recognition. Face and voice recognition system, it will also receive information about the success of
can be used to identify a person so that the social robot can its strategies and, in the best case, be able to adapt them it-
remember a customer or even, in conjunction with appropri- self.
ate data, name them. It may also be used to determine age
and gender, which can be important in the sales process. Should Social Robots Manipulate Customers?
Emotion recognition provides information about the state of
the customer upon entering the store, while shopping and Nowadays, social robots and especially humanoid variants
receiving advice, and finally when leaving the store. are being used for several purposes in the education system,
With the help of such means, it is also possible to catego- therapy and care, entertainment, hotel business, and retail
rize and select customers. It is possible to ask them about sectors (Alves 2021). In doing so, these robots are becoming
their wishes during a conversation and then make corre- increasingly intelligent and their “behavior” less and less
sponding suggestions, or to classify and assign them on the distinguishable from humans’ behavior, making them well
basis of the automatically recognized age, gender, body suited for consulting and sales assistance in retail stores.
shape, and state of mind, for example to relevant discount Perhaps it is only a matter of time before the technology
campaigns and special offers. If the robot were mobile or will reach a point when people will no longer deal with hu-
movable, it could take the customer to the checkout or at man advisors and sellers in retail stores but with humanoid
least show them the way – for security reasons, most social robots – if stationary retail still has a chance at all. This is
robots are found in a static (additionally secured) location in also supported by the advantages of automation listed at the
the retail store or shopping mall. beginning. Of course, there are also disadvantages, such as
the lack of genuine social contact. However, according to
several studies, robots are expected to be able to advise and
10
sell at least as well or even better than human beings by for industries in general. Moreover, most of the guidelines
2025. It is expected that they will be more empathetic, more and regulations are vaguely defined, so there are many ways
situation-specific, more flexible, more sophisticated, and to circumvent them and there are no subsequent sanctions if
more versatile. In short: more successful (Scheible 2019). one does not adhere to the guidelines.
Knowing that human advisors and salespeople can ma- For SQ3, the goal was to find out if ethical guidelines and
nipulate customers, even though an international code of policies were established about who must perform the final
ethics for sales and marketing (Sirgy 2014) exists, albeit inspection of the robots before they are placed into service.
very succinctly and not everyone adheres to it, the question In this aspect, partially created ethical guidelines and poli-
arises whether a social robot could also manipulate custom- cies were also disclosed. However, these tests or inspections
ers in retail stores to obtain an advantage. Manipulation here do not explicitly refer to social and humanoid robots in retail
means that it directs their intentions so that, in extreme stores. It is not specified which tests or inspections should
cases, they buy something they don’t even want or need. be performed, and it is not mentioned who should or must
complete the final tests or inspections before deploying ro-
Study about Manipulative Robots bots in retail stores. Once again, the guidelines and regula-
In her master thesis at the School of Business FHNW, Lili- tions are very vaguely defined, and no sanctions exist.
ana Alves conducted a study on manipulative consulting and With SQ4, the aim was to find out how potential custom-
sales robots. The purpose of the study was both to provide ers in shopping malls and stores react, what they think and
transparency in the area of negative manipulation by human- feel, when confronted with a manipulative humanoid advi-
oid robots and to fill the gap in ethical considerations of cus- sor or sales robot in the retail sector. Thereby, it was identi-
tomer manipulation by humanoid consulting and sale assis- fied that different thoughts, feelings, and reactions exist to-
tance robots in retail stores (Alves 2021). wards manipulative robots in retail stores. In fact, some peo-
The main research question (RQ) was to determine ple have an utterly negative opinion towards manipulative
whether it is ethical to intentionally program humanoid con- robots, others are neutral, and others are even positive. Yet,
sulting and sales robots with manipulation techniques to in- it can be concluded that generally, people do not want to be
fluence the customer’s purchase decision in retail stores. manipulated by humanoid robots and would therefore rather
Moreover, to answer this central question, five sub-ques- avoid these kinds of robots in the future or be more cautious
tions (SQ) were defined and answered based on an extensive when interacting with them.
literature review and a survey conducted with potential cus- Lastly, SQ5 should find out if potential customers ac-
tomers of all ages and varying socio-demographic charac- cepted a manipulative and humanoid advisor or sales robot
teristics. For SQ1, the goal was to find out how humanoid in a retail store. In this context, it was investigated whether
consulting and sales robots can manipulate customers in re- customers might accept a manipulative robot in a retail
tail stores. Thereby, it was identified that social and human- store, but only if the manipulation is used positively by en-
oid robots can be programmed with manipulation content hancing the customer’s well-being or shopping experience.
and are technically capable of manipulating customers, sim- If, on the other hand, the manipulation is used to negatively
ilar to human advisors or salespeople. As already indicated, influence the customer, it becomes neither acceptable nor
there are several ways to do this, namely through vocal ethically justifiable. The survey that produced these findings
pitch, pacing of speech, voice volume, sentence melody, ar- on SQ4 and SQ5 will now be looked at in more detail.
ticulation, tone, words, semantics (e.g., questioning tech-
niques, content-based manipulation concepts, argumenta- Online Survey on Robot Manipulation
tion concepts etc.), linguistic particularities, technical terms, The co-author conducted an online survey between Febru-
foreign languages, posture, movement, gestures, facial ex- ary 9 and March 14, 2021. It was accessed 751 times,
pressions, and robot enhancement. whereof 328 participants completed the survey (completion
SQ2 aimed to determine if there are already ethical guide- rate of 43.8 %). Approximately two-thirds were male, one-
lines and policies to prevent humanoid robots from manipu- third female. All age groups were represented, with the larg-
lating customers’ purchasing decisions in the retail sector est numbers being 26- to 35-year-olds (109), 36- to 45-year-
that developers and robot users must adhere to. Here, it was olds (79), and 46- to 55-year-olds (66), followed by 56- to
determined that manipulation is a known issue among vari- 65-year-olds (41) (Alves 2021). In the evaluation of the sur-
ous leading players (e.g., professional associations, national vey no distinction was made between gender or age.
federations, and industry) in Europe, which is why some eth- Two practical cases were presented in the online survey
ical guidelines and principles have already been created to in order to answer SQ4 and SQ5. In the first, Pepper directly
avoid manipulation. However, manipulation is not explicitly informs a customer that it has manipulated them. In the sec-
related to social and humanoid robots, and these guidelines ond, it does not inform the customer directly, and the cus-
have not been developed specifically for retail sectors but tomer discovers in an indirect way that they have been
11
manipulated by the robot. In both cases, the assumed manip- being used in the customers’ interests. Other participants in-
ulation occurred through a specific form of content and tac- dicated that they would even celebrate the robot, congratu-
tical strategies that are also often used by human consultants late it and find the situation amusing. The participants men-
and salespeople. Overall, it is primarily design, dialogue ca- tioned that they would be grateful to learn to be more careful
pability, and information transfer that are affected. in the future and listen more to their intuition.
In the first case, around 108 people would be surprised Most participants would probably want to return the pur-
that they have been manipulated, 93 people would feel de- chase and receive a refund. Most would share their experi-
ceived or betrayed, and 91 people would be upset. In the ences with their personal and professional environment
second case, 143 people said they would feel deceived or (e.g., via social media). Some participants mentioned that
betrayed, 139 people would be upset, and 93 people would they would avoid such social robots in the future. Others
be surprised. In practice, the second case would be most would probably not enter the same retail store or never seek
likely to occur in real life scenarios. These two cases illus- advice from a system like that again. However, other partic-
trate that customers would most probably feel deceived or ipants noted that they would simply be more careful when
betrayed and upset when being manipulated by a robot. interacting with a robot in the future.
Summarizing the survey’s main findings, the majority of What was clear was that most participants do not want
the 328 participants prefer to be advised by a human advisor and do not accept manipulative robots in retail stores and
in a retail store (Alves 2021). Generally speaking, spontane- further believe that society should not accept them either.
ous advice is accepted, but this depends on the situation and For most people, it is not ethically justifiable to use manip-
the cordiality of the advisor or salesperson. Around 260 par- ulative robots in this context. According to these partici-
ticipants have no prejudices against human advisors or pants, robots in retail stores should be regulated for the pur-
sellers. However, those who have prejudices believe that hu- pose of “negative manipulation”, but not banned from oper-
man advisors or sellers only want to sell without focusing ating altogether. This is merely to say that the harmful facets
on the customer’s needs. Namely, they mainly aim to fulfil of manipulation should be regulated. Instead, it is more im-
their own sales quota and sales targets, and to maximize portant that a society becomes educated about such harmful
profits to earn additional commissions and bonuses. A fur- actions to make responsible decisions.
ther prejudice is the lack of knowledge of the products or Here, too, not all participants saw the situation the same
services they sell. Participants expect or believe that they way. For a smaller number it seemed appropriate for such
often experience manipulation by humans during the retail manipulative robots to be used in retail stores. In fact, they
store’s counselling or sales process. Thereby, they feel ei- stated that these robots are morally acceptable and should
ther negatively or neutral when they become aware that they also be approved by society because they perform the same
have been manipulated. actions and tactics as humans. Thus, it is the customer’s
More than half of the participants are aware of the robot choice whether to be manipulated by them or not. As long
Pepper, and that manipulative robots exist in theory or prac- as humans are allowed to influence customers negatively,
tice. However, so far, only a small number of people (56) robots should be allowed to do it, too, primarily because hu-
have had direct interaction with Pepper, and among these mans program them. For this reason, in these participants’
only ten people have ever received a consultation from it. opinion, these robots are ethically justifiable. Overall, the
Up to now, the experience with Pepper was generally rated participants still prefer human interactions and would rather
as neutral by the participants. avoid the humanoid robot in retail altogether.
When the survey confronted the participants with two hy-
pothetical cases in which a manipulative robot negatively Interim Conclusion
influenced them, about half of the participants stated that The study, not relying only on the survey, can answer the
they had not expected such a situation and were initially main research question (RQ) as follows (Alves 2021): It is
somewhat surprised, amazed, and speechless. They would neither ethical for software developers to program robots
primarily doubt themselves as well as their purchase, and with manipulative content nor is it ethical for companies to
subsequently, they would probably feel angry, irritated, de- actively use these kinds of robots in retail stores to system-
ceived, cheated, and unpleasant. atically manipulate customers in order to obtain an ad-
However, other participants viewed this neutrally and as- vantage. Business is about reciprocity, and it is not accepta-
sumed that manipulation can happen anywhere. In this con- ble to systematically deceive, exploit, or manipulate cus-
text, the participants believe that it depends on the customer tomers to attain any kind of benefit.
whether they make the purchase, since there is no force ap- However, it turns out that some survey participants find a
plied. Ultimately, the robots are programmed by humans, manipulating robot acceptable or at least entertaining and
which should make situations like this predictable. Further- amusing. Some also believe that humanoid or social robots
more, both robots and humans have the same task to fulfil, should be allowed to cheat as long as humans do. This result
and it would be naive to believe that the robots are only
12
must be taken seriously. Even if one were to find a different Ethical Perspective
overall picture in further (larger and more representative) In principle, it can be considered whether a social robot or a
surveys, the individual statements would still remain. They service robot in retail must have a humanoid design. When
will not be overly considered in the following section but Pepper looks at the customer with its big eyes, it generates
they will be taken into account. emotions in him or her, which it can recognize and reflect in
its behavior and speech (see Fig. 3). In this context, one can
Recommendations for Companies certainly imagine a capacity for deception and fraud. At
least users are made to believe that they have something
The following section provides recommendations for com- alive in front of them, and they are manipulated in a certain
panies looking to deploy social robots in retail settings. way as one exploits their evolutionary tendency to react to
Some recommendations are of a general nature or are de- something alive in a particular manner.
rived from previous experience in this area, while others are The operator should safeguard the use of social robots via
based on the findings of this paper and in particular on the ethical guidelines, similar to what the developer may have
evaluation of the study. done previously during programming. These can be adopted
from relevant initiatives and government agencies (in Eu-
Technical Perspective rope, the High-level expert group on artificial intelligence
It is fundamentally a decision for a company to employ a should be mentioned, see Veale 2020) but should be adapted
robot that supplements or replaces an employee. It is also to the operator’s own practice. It is important that the ethical
fundamentally their decision to use a classic service robot guidelines are concrete, i.e., useful and implementable. In
without or with only a few social skills or a social robot with addition, non-compliance should result in sanctions. Legal
service skills. In doing so, the possibilities of AI can also be provisions should also be taken into account, especially with
considered to a greater or lesser extent. regard to data protection and transparency.
From a technical perspective, it is first important to ensure In particular, the bias discussion should be considered
the functionality and security of the systems, i.e., the robot during development and operation. The social or humanoid
itself and the systems connected to it, such as databases and robot should not show any prejudices toward customers, just
AI systems used for face recognition. This serves to estab- like consultants and salespeople. From this, there are behav-
lish fundamental trust in this type of technology and in this iors that would be fundamentally prohibited, such as, for in-
form of service. stance, negative statements made in the advisory and sales
Furthermore, a solid database must be guaranteed. The conversation with regard to age, gender, and ethnicity.
robot should know all the products in question and be able Prejudices and distortions should also be avoided through
to name the prices and discounts correctly. Knowledge the use of data. This point is related to the technical perspec-
about the company itself and its customers is also important. tive. The data and algorithms must be checked for biases,
When it comes to “world knowledge”, many conversational and actual biases must be removed. However, this will not
agents and also social robots access Wikipedia, although it always be possible, and correlations can prove useful for
is not always reliable. Here, too, an alternative should be grouping and identifying potential customer preferences.
considered, even if it is merely Wikipedia articles that have That many customers do not want to be manipulated must
been additionally reviewed (by the company’s own experts). be taken into account. In this respect, it would tend to be
It is also possible to offer customers various technical easier to implement this in a robot than a consultant or sales-
choices. For example, at the beginning of the consultation, person, which is again related to the technical perspective.
they could select via a menu on the display whether the so- However, some respondents in the study stated that they had
cial robot should act more in the guise of a neutral sales sys- no problem with being manipulated, partly because this
tem or that of a salesperson in the spirit of the MOME, the could contribute to their shopping experience. This could
morality menu (Bendel 2020). Other aspects, such as the also be covered by a choice option, which again has to do
voice (female, male, or neutral) and the personality (serious, with the technical perspective.
casual, funny, etc.) could also be selected. Some survey participants seem to think that social robots
Last but not least, it would be possible to let the user select should be allowed to manipulate when humans are. This
or limit the AI-based systems individually. He or she could, could be from a kind of sense of justice, although robots and
for instance, do without facial recognition and related emo- humans are fundamentally different entities (which is not
tion recognition, thus protecting his or her privacy and in- necessarily recognized by the users, especially since they
formational autonomy. However, this would entail accept- are social robots that deliberately blur the differences). It
ing restrictions in the shopping experience. They must also was noted, however, that it is the programmer (together with
be prepared to be informed about the opportunities and risks, other parties) who commands the manipulation, so to speak.
and be able to understand them. The robot is simply reproducing human reality.
13
It could be the case that despite the neutral design or the It is also worth asking whether certain manipulations are
possibility of choice on the part of the customers, manipula- not simply part of the business – what some of the interview-
tions of the robot arise, whether intended or unintended. If, ees believed. Advertising makes many promises that can
despite assurances to the contrary, certain sales tricks and hardly be kept, but which promote sales, and certain strate-
attempts at outwitting were used, this would be considered gies of the sales staff also increase sales. It has become a
a breach of trust and critical from an ethical perspective. game whose limits are constantly being tested. Just as some
However – as the survey also revealed – a robot that is not like to watch advertising that uses exaggeration, some might
trustworthy could, in a certain sense, be an opportunity. Be- like to face social robots that try to trick them.
cause if we become suspicious, we are less likely to fall for However, too much manipulation, whether it comes from
tricks of all kinds. This principle of experience also applies the consultant or salesperson or from the social robot, could
to technical systems. Responsible and trustworthy artificial result in the original goal of reinforcing physical locality be-
intelligence is an important goal but educating users to its ing reversed. Those who do not feel comfortable in the store,
opposite seems to be just as important. who are deceived and cheated, will ultimately prefer e-com-
Last but not least, it must be noted that a store or shopping merce – or send others to do the shopping.
mall is not particularly the place for unembellished truth. It
contributes to a customer’s well-being to be flattered and
complimented. Incidentally, he or she also does not want to
be constantly told the truth about his or her appearance or
behavior outside of shopping, especially if he or she does
not conform to social expectations in their appearance.
Economic Perspective
By using humanoid or social robots, a company can try to
revitalise or increase the attractiveness of brick-and-mortar
retail. It can demonstrate its willingness to transform itself,
its business, and its innovative strength, thus setting itself
apart from the competition. That being said, the novelty ef-
fect could quickly wear off.
A stationary retail company benefits from automation and
AI methods. It can place its workforce in other areas or lay
off workers, thereby reducing personnel costs (which raises
ethical questions) and get to know its customers better. It
can build a valuable data base based on the conversations
and (re-)actions of its customers to better advise and assist
groups or individuals, even those who do not fit the standard
customer type.
Admittedly, some advisors and salespeople will see the
social robot as direct competition, replacing them in their
core activities. Meyer at al. surveyed “frontline employees”
(FLEs) in a study: “The findings extend prior studies on
technology acceptance and resistance and reveal […] that
FLEs perceive service robots as both a threat and potential
support. Moreover, they feel hardly involved in the co-crea- Fig. 3: Pepper (Westfield, 2017)
tion of a service robot, although they are willing to contrib-
ute.” (Meyer et al. 2020b) The economic perspective is actually a much broader one.
If a company used manipulative social robots, this could When retailers employ social robots that obviously deceive
damage its reputation. Even if it designed the system neu- and cheat, this shapes the image of other social robots that
trally or granted a choice, it is not immune to criticism, es- are urgently needed in service or care and therapy. This
pecially if it turns out that manipulations were nevertheless harms their manufacturers and the organizations and indi-
present. Competitors can strike back here, so to speak, espe- viduals who rely on their use. From this point of view, it
cially those retailers who reject the use of social robots for would be economically and ultimately also ethically advis-
certain reasons, e.g., because they value social contacts. able to keep manipulation in this area low.
14
Summary and Outlook Bendel, O. 2020. The Morality Menu Project. In Nørskov, M.;
Seibt, J.; Quick, O. S. (eds.) Culturally Sustainable Social Robotics
It can be assumed that social robots in the form of consulting – Challenges, Methods and Solutions: Proceedings of Robophilos-
ophy 2020. IOS Press, Amsterdam, 257–268.
and sales robots will play an increasingly important role in
retail in the future as their capabilities grow because they are Bendel, O. (ed.) 2018. Pflegeroboter. Wiesbaden: Springer Gabler.
attractive to customers, the advisory service can be individ- Bendel, O.; Schwegler, K.; and Richards, B. 2017. Towards Kant
Machines. In The 2017 AAAI Spring Symposium Series. AAAI
ualized and simultaneously standardized, and the sales pro-
Press, Palo Alto 2017, 7–11.
cess leading to a deal can be at least partially automated.
Coursey, K. 2020. Speaking with Harmony: Finding the right thing
These developments call for a scientific and sales-focused to do or say ... while in bed (or anywhere else). In Bendel, O. (ed.)
practical examination of buyer-robot interactions. In this Maschinenliebe: Liebespuppen und Sexroboter aus technischer,
context, the optimization of movement and natural language psychologischer und philosophischer Sicht. Wiesbaden: Springer
capabilities are central. The social robot is at the customer’s Gabler, 35–51.
side, in dialogue with them, listening to their questions and Evangelista, B. 2016. Robots greet Westfield mall shoppers in San
giving answers in different languages. Francisco, San Jose. SFGATE, 22 November 2016.
https://www.sfgate.com/business/article/Robots-greet-Westfield-
If social and specifically humanoid robots like Pepper
mall-shoppers-in-San-10631291.php.
(whose production was discontinued in 2020, which raises
Kelly, T. 2020. Japanese robot to clock in at a convenience store in
certain questions), NAO, Paul, and Cruzr have found their test of retail automation. 15 July 2020. In Reuters.com,
way into shopping stores and malls to attract, greet, and wel- https://www.reuters.com/article/us-japan-tech-robot-
come customers, to inform, advise, and in the future per- idUSKCN24G138.
suade them to buy – should they behave like human advisors Meyer, P.; Jonas, J. M.; and Roth, A. 2020a. Exploring customer’s
and salespeople, i.e. manipulate customers? Or should they acceptance of and resistance to service robots in stationary retail –
be more honest and reliable than humans are? This article a mixed method approach. In ECIS 2020 Proceedings at AIS Elec-
tronic Library (AISeL). Research Papers, 9,
explored this question.
https://aisel.aisnet.org/ecis2020_rp/9.
The conclusion is that manipulative behavior will please
Meyer, P.; Jonas, J. M.; and Roth, A. 2020b. Frontline Employees’
a small number of customers. They are interested in the Acceptance of and Resistance to Service Robots in Stationary Re-
game and the ambiguity or see a kind of equality in this pos- tail: An Exploratory Interview Study. SMR Journal of Service
sibility. The majority of customers, however, are likely to be Management Research, Volume 4, 1/2020:21–34.
interested in not being manipulated, and in a robot ulti- Prasad, V.; Stock-Homburg, R.; and Peters, J. 2021. Human-Robot
mately being less manipulative than a human advisor or Handshaking: A Review. In arXiv.org, 14 February 2021,
salesperson. Ultimately, fair, honest, and trustworthy robots https://arxiv.org/abs/2102.07193.
in retail are a win-win for everyone, not least for the com- Scheible, K.-G. 2019. Roboter schlägt Mensch – Verhandlungen
pany and yet, customers can be given choices that serve their der Zukunft. In Buchenau, P. (ed.) Chefsache Zukunft: Was Füh-
rungskräfte von morgen brauchen. Wiesbaden: Springer Gabler,
usual or desired individual shopping experience. So how fair 507–520.
is fair? This depends on the wishes of the customers, but
Sirgy, M. J. 2014. Real Estate Marketing: Strategy, Personal Sell-
most of them expect to be served honestly and transparently ing, Negotiation, Management, and Ethics. London: Taylor &
by a social robot. Francis.
Song, Y.; and Luximon, Y. 2021. The face of trust: The effect of
robot face ratio on consumer preference. Computers in Human Be-
References havior, Volume 116, March 2021, 106620. https://www.sciencedi-
rect.com/science/article/pii/S0747563220303678.
Aaltonen, I.; Arvola, A.; and Heikkilä, P. 2017. Hello Pepper, May
I Tickle You?: Children’s and Adults’ Responses to an Entertain- Veale, M. 2020. A Critical Take on the Policy Recommendations
ment Robot at a Shopping Mall. In HRI ’17: Proceedings of the of the EU High-Level Expert Group on Artificial Intelligence. Eu-
Companion of the 2017 ACM/IEEE International Conference on ropean Journal of Risk Regulation:1–10.
Human-Robot Interaction, March 2017, 53–54. Vontobel, N.; and Weinmann, B. 2017. So motzt die Migros Kon-
Alves, L. 2021. Manipulation by humanoid consulting and sales sumtempel im Kampf gegen Onlineshops auf. Tagblatt, 7 October
hardware robots from an ethical perspective. Master Thesis. Olten: 2017. https://www.tagblatt.ch/wirtschaft/so-motzt-die-migros-
School of Business FHNW. konsumtempel-im-kampf-gegen-onlineshops-auf-ld.1456998.
Bendel, O. (ed.). 2021a. Soziale Roboter: Technikwissenschaftli- Werdayani, D.; and Widiaty, I. 2021. Virtual fitting room technol-
che, wirtschaftswissenschaftliche, philosophische, psychologische ogy in fashion design. In IOP Conf. Ser.: Mater. Sci. Eng. 1098
und soziologische Grundlagen. Wiesbaden: Springer Gabler. 022110.
Bendel, O. 2021b. Das steht Ihnen aber gut!: Empfangs-, Bera- Woollaston, V. 2016. The phone store run by ROBOTS: Tokyo
tungs- und Verkaufsroboter im Detailhandel. In Bendel, O. (ed.) firm replaces staff with a team of Pepper the ‘emotional’ human-
Soziale Roboter: Technikwissenschaftliche, wirtschaftswissen- oids. Mail Online, 24 March 2016. https://www.dai-
schaftliche, philosophische, psychologische und soziologische lymail.co.uk/sciencetech/article-3507541/The-phone-store-run-
Grundlagen. Wiesbaden: Springer Gabler, 495–515. ROBOTS-Tokyo-firm-replaces-staff-10-versions-Pepper-emo-
tional-humanoid.html.
15