=Paper=
{{Paper
|id=Vol-3812/paper10
|storemode=property
|title=Responsible Configuration Using LLM-based Sustainability-Aware Explanations
|pdfUrl=https://ceur-ws.org/Vol-3812/paper10.pdf
|volume=Vol-3812
|authors=Sebastian Lubos,Alexander Felfernig,Lothar Hotz,Thi Ngoc Trang Tran,Seda Polat-Erdeniz,Viet-Man Le,Damian Garber,Merfat El-Mansi
|dblpUrl=https://dblp.org/rec/conf/confws/LubosFHTELGM24
}}
==Responsible Configuration Using LLM-based Sustainability-Aware Explanations==
https://ceur-ws.org/Vol-3812/paper10.pdf
Responsible Configuration Using LLM-based
Sustainability-Aware Explanations
Sebastian Lubos1,* , Alexander Felfernig1,* , Lothar Hotz2 , Thi Ngoc Trang Tran1 , Seda Polat-Erdeniz1 ,
Viet-Man Le1 , Damian Garber1 and Merfat El-Mansi1
1
Institute of Software Technology, Graz University of Technology, Graz, Austria
2
Hamburger Informatik Technologie-Center e.V., Hamburg, Germany
Abstract
Configuration systems play an important role in achieving the sustainable development goals (SDGs) defined by the United Nations. As
decision support systems, configurators help users to decide which components or features to include in or exclude from a configuration.
An important task of configurators is the provision of explanations which help to achieve goals such as increasing configuration
understandability, increasing a user’s trust, and persuading users/customers to include specific configuration components. Our goal in
this paper is to introduce the concept of „sustainability-aware explanations“ which can help to support the sustainable development goals.
The type of explanations we propose in this context are somehow orthogonal to typical explanations used in industrial configuration
environments. A major objective in this context is to follow a „less-is-more“ principle focusing on different aspects of the idea of
„responsible configuration“ which refers to configuration techniques explicitly supporting the mentioned sustainability goals. We report
the initial results of an evaluation that provide insights on potential impacts of the proposed explanations.
Keywords
Explanations, Sustainability, Green Configuration, Responsible Configuration, Configuration for Good, Nudging, Persuasion, Knowledge-
based Configuration
1. Introduction relevant user requirements that lead to the determination of
a specific configuration. Furthermore, why not explanations
The 17 sustainable development goals (SDGs) defined by focus on supporting users in situations where no solution
the United Nations (UN) provide a blueprint for peace and can be identified [13, 14, 15]. From the application point
prosperity on our planet.1 Examples of such goals are of view, explanations can be applied to achieve different
good health and well-being (e.g., in terms of fostering the goals [16].2 Examples thereof are efficiency (reducing the
consumption of healthy food), responsible consumption and time that is needed to complete a configuration task),
production (e.g., in terms of reduced energy consumption), persuasiveness (convincing users to change their component
and sustainable cities and communities (e.g., in the context selection behavior), transparency (making the inclusion or
of tourism, avoiding negative environmental impacts and exclusion of specific components transparent to the user),
taking into account the local communities and cultural trust (increasing a user’s confidence in the configuration
heritage) [1]. system), scrutability (making it possible for the user to
Knowledge-based configuration [2, 3, 4, 5] can be adapt the configurator behavior, e.g., in terms of the used
regarded as a core-technology of mass customization component inclusion/exclusion strategy), and satisfaction
[6]. On the basis of configurators, users are enabled to (e.g., increasing the usability of a configuration system).
design a product in an individualized fashion that fits These goals must be regarded as examples – for related
their wishes and needs. In configuration settings, we details we refer to [11, 16, 17, 18].
can observe an ever-increasing demand for taking into In this paper, we focus on the persuasion aspect of
account sustainability aspects [7, 8]. Following the basic explanations [19]. More precisely, we analyze possibilities
definition of „configuration“ given by Sabin and Weigel [3], to formulate explanations in such a way that users are
i.e., „configuration is a special case of design activity where nudged towards more sustainability-aware configuration
the configured artifact is assembled from a fixed set of well- decisions. Following a „less-is-more“ principle, we show
defined component types and components are interacting in how to formulate explanations following Cialdini’s six
predefined ways“, we define „responsible configuration“ as principles of persuasion [20] (see Table 1).
„configuration which takes into account the United Nation’s Sustainability-aware explanations have to focus on
sustainable development goals“. argumentations including sustainability aspects. Our
In knowledge-based systems, explanations can be applied formulation of such explanations is based on large
for different purposes [9]. First, so-called why explanations language model (LLM) prompts [21] which help to
[10, 11, 12] focus on the aspect of mentioning the most associate sustainable development goals with the mentioned
persuasive principles. For example, in the context of
ConfWS’24: 26th International Workshop on Configuration, Sep 2–3, 2024, car configuration, explanations could refer to the positive
Girona, Spain environmental aspects of purchasing smaller cars or on the
*
Corresponding author.
$ slubos@ist.tugraz.at (S. Lubos); afelfern@ist.tugraz.at (A. Felfernig);
advantages of electric vehicles compared to gasoline-driven
lothar.hotz@uni-hamburg.de (L. Hotz); ttrang@ist.tugraz.at ones.
(T. N. T. Tran); spolater@ist.tugraz.at (S. Polat-Erdeniz); Positive impacts of such sustainability-aware
vietman.le@ist.tugraz.at (V. Le); dgarber@ist.tugraz.at (D. Garber); explanations can be, for example, higher-quality
merfat.el-mansi@student.tugraz.at (M. El-Mansi) configuration decisions, a lower amount of unneeded
0000-0002-5024-3786 (S. Lubos); 0000-0003-0108-3146 (A. Felfernig);
components, and components with less negative
0000-0001-7370-7726 (L. Hotz); 0000-0002-3550-8352 (T. N. T. Tran);
0000-0001-5778-975X (V. Le)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License 2
Attribution 4.0 International (CC BY 4.0). The categorizations of [11, 16] have been developed in the context of
1
https://sdgs.un.org/goals recommender systems but can also be applied in configuration contexts.
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
Table 1 Scenario 2: Long vs. standard range battery. The idea
Cialdini’s principles of persuasion [20]. is to make configurator users interested in purchasing a car
principle semantics with a long-range battery aware of the sustainability aspects
feeling of an obligation to give something of standard-range batteries. To support such explanations,
reciprocity
back we have generated LLM-based explanations using the
reduced item availability increases following LLM prompt: person A wants to purchase an
scarcity
preparedness to purchase electric car and is interested in a long-range battery. Please
experts have an increased influence on provide persuasive arguments against purchasing a long range
authority
users
battery following the six persuasion principles of Cialdini.
commitment
and
users prefer to be consistent with their The corresponding LLM-generated sustainability-aware
articulated preferences explanations are depicted in Table 3.
consistency
users like to comply with other users who
liking
are similar to themselves
Scenario 3: Car not needed in city center.
users follow the opinions (of a
social proof
representative set) of other users
Configurator users should think about the advantages of
not having a car when living in the city center. We have
generated related LLM-based persuasive explanations
environmental impacts [8]. From a commercial point of using the following LLM prompt: person A who lives
view, such explanations might appear – at least to some directly in the city center with various connections to public
extent – counterproductive due to potential consequences transportation wants to purchase a car. Please provide
in terms of decreasing turnovers. Thus, sustainability- persuasive arguments against purchasing a car following
aware explanations are often in contrast to explanations in the six persuasion principles of Cialdini. The corresponding
mainstream configuration environments which focus on sustainability-aware explanations are depicted in Table 4.
increasing sales rates in most of the cases.
The contributions of this paper are the following. First, Scenario 4: Less costly car due to financial situation.
we propose the concept of sustainability-aware explanations The idea is making configurator users with limited financial
for configurations. Second, we provide reference examples resources intending to purchase an expensive car to
of such explanations in the automotive domain. Third, we change their mind and purchase a less expensive car.
present initial results of a corresponding evaluation. To support such explanations, we have generated LLM-
The remainder of this paper is organized as follows. In based explanations using the following (example) LLM
Section 2, we provide different examples of LLM-generated prompt formulation: person A with very limited financial
sustainability-aware explanations in the car configuration resources and a family with three children wants to purchase
domain. Thereafter, we discuss initial results of a related an expensive car. Please provide persuasive arguments
evaluation (Section 3). In Section 4, we discuss threats to against purchasing an expensive car following the six
validity. Finally, we conclude the paper with Section 5. persuasion principles of Cialdini. The related LLM-generated
explanations are depicted in Table 5.
2. Sustainability-Aware Explanations
3. Evaluation
with LLMs
Properties of LLM-based explanations. In Table 6,
In the following, we discuss scenarios where sustainability- we summarize the different argumentation lines generated
aware explanations can have an impact on user decisions. by the large language model (LLM). (1) In the context
All scenarios are related to car configuration where users of the persuasion dimension reciprocity, LLM-generated
receive explanations of current configurations. The major explanations refer to the aspect of „giving something
goal of such explanations is to make users think about their back to the community“, for example, purchasing an
current configuration settings and to potentially adapt their eco-friendly vehicle can be a way of giving back to the
articulated preferences. Consequently, our explanations are environment. (2) Explanations related to the persuasion
not in the line of why or why not explanations but focus dimension scarcity on the one hand refer to decreasing
more on indicating potential alternatives to the current incentives for sustainable equipment (e.g., cars), on the
configuration, i.e., a kind of why not choose something else other hand to limited resources (e.g., financial resources).
explanation. All example explanations in this paper have (3) In the context of the persuasion dimension authority,
been generated on the basis of the LLM ChatGPT 3.5.3 different experts such as environmental experts are used as
representatives of authorities. (4) Explanations related to
Scenario 1: SUV vs. smaller car. The idea is to make the persuasion dimension commitment assume an existing
persons (configurator users) who intend to purchase an commitment of the current user, for example, already proved
SUV more aware of sustainability aspects of smaller cars. in previous configuration sessions. (5) In the context of the
To support this, we have generated LLM-based explanations persuasion dimension liking, LLM-generated explanations
using the following (example) LLM prompt: Assume the refer to a user’s family, friends, and neighbors (e.g., your
following scenario: person A wants to purchase a car and is family will like your decision). (6) Explanations related to
interested in an SUV. Please provide persuasive explanations the persuasion dimension social proof refer to trends of
against purchasing an SUV following the six persuasion peers, the wider community, and similar families.
principles of Cialdini. The resulting explanations are
depicted in Table 2. LLM-based impact estimates. Using the LLM prompt
which of these explanations would have the highest persuasive
3
https://chat.openai.com impact on a user? Please provide a ranking., we asked the
Table 2
Sustainability-aware explanations related to the aspect of purchasing a smaller car compared to an SUV.
principle LLM explanation
Consider the environmental impact and how it affects everyone. Choosing a more eco-friendly vehicle can
reciprocity be your way of giving back to the community and environment. Others will likely appreciate and
reciprocate your thoughtful decision, fostering a positive cycle of environmentally conscious choices.
commitment You’ve always been committed to sustainability and reducing your carbon footprint. Opting for an SUV
and contradicts these values due to its higher emissions and fuel consumption. Staying consistent with your
consistency commitment to the environment means choosing a more fuel-efficient and eco-friendly car.
Look at the growing trend among your peers and the wider community towards hybrid and electric vehicles.
social proof Many people are switching to these options because they are better for the environment and more
economical in the long run. Following this trend shows you’re in tune with progressive, responsible choices.
Environmental experts and automotive industry leaders are advocating for smaller, more efficient vehicles to
authority combat climate change and reduce pollution. Following the guidance of these authorities can help you make
a well-informed decision that aligns with expert recommendations.
Your friends and family who care about the environment will appreciate and support your choice of a more
liking sustainable vehicle. This decision can strengthen your relationships, as they see you aligning your actions
with shared values and concerns.
With increasing regulations and shifts towards renewable energy, fuel-efficient and eco-friendly vehicles
scarcity may soon become the norm, potentially leading to fewer incentives for traditional gas-guzzling SUVs. Take
advantage of the current incentives for hybrid or electric vehicles before they become more scarce.
Table 3
Sustainability-aware explanations related to the aspect of purchasing a standard compared to a long-range battery.
principle LLM explanation
Think about the balance between what you get and what you give. By opting for a standard-range battery,
you’re reducing the environmental impact associated with producing larger batteries. This considerate
reciprocity
choice benefits the community and environment, and others may appreciate your mindful decision, leading
them to reciprocate with similar sustainable choices.
You’ve been committed to reducing your carbon footprint and promoting sustainability. A standard-range
commitment
battery is more aligned with these values, as it requires fewer resources to produce and has a lower
and
environmental impact. Staying consistent with your commitment to sustainability means choosing the
consistency
option that minimizes ecological harm.
Many environmentally conscious drivers are opting for electric vehicles with standard-range batteries
because they strike a good balance between range and environmental impact. By choosing a standard-range
social proof
battery, you’re aligning with a growing community of eco-friendly drivers who prioritize sustainability over
convenience.
Experts in environmental science and sustainable technology often highlight the advantages of smaller,
more efficient batteries. They point out that while long-range batteries offer more convenience, the
authority
production and disposal of larger batteries have a higher environmental cost. Following the advice of these
authorities can help you make a more environmentally responsible decision.
Your friends and family who are also concerned about environmental issues will likely appreciate your
liking choice of a standard-range battery. This decision can enhance your relationships as they see you making
sacrifices for the greater good and staying true to your shared values.
As technology advances, the efficiency of electric vehicle batteries is continually improving. Standard-range
batteries are becoming more capable, offering longer ranges and better performance. By choosing a
scarcity
standard-range battery now, you can take advantage of current technologies and innovations without
feeling the pressure to invest in a long-range battery that may soon become less necessary.
LLM also for a ranking of the impact of the generated between the explanations ranked highest by the LLM and
explanations following the idea of LLM self-evaluation those selected by the study participants. In this context,
[22]. In most settings, the dimensions authority and the authority principle has been regarded as relevant in all
commitment & consistency have been regarded as the two example configuration scenarios. This initial result indicates
most impactful persuasion dimensions (see Table 7). The basic LLM capabilities to recommend persuasion strategies.
related LLM-based argument is that the latter appeals
directly to a person’s desire to act in accordance with their
past beliefs and statements and the former has a high impact 4. Threats to Validity
specifically in scenarios where the underlying topic involves
The explanation concepts presented in this paper are based
technical details. Finally, the dimension of social proof can
on one selected configuration scenario (car configuration).
be impactful in situations where social trends and peer
These explanations are just high-level examples and many
behaviors impact decisions.
further (also more detailed ones) can be envisioned for
car configuration (e.g., the sustainability aspects of a less
Feedback of study participants. For a very first powerful car engine) and beyond. In our work, we did not
evaluation beyond the LLM feedback, we conducted a small focus on a specific phase of a configuration process, i.e.,
user study in our research group with N=10 participants the mentioned explanations could even be used before the
(computer science students at our university). The configuration process has been started (e.g., as explanations
participants were asked to rank the different explanations in wish lists or product information). Gaining more
according to their potential persuasive impact. The results related insights is a major focus of our future research.
are included in Table 7. Overall, there exists an overlap
Table 4
Sustainability-aware explanations related to the aspect of waiving a car in the city center.
principle LLM explanation
Consider how much the city has invested in developing a robust public transportation system. By
using these resources instead of buying a car, you show appreciation and reciprocate the city’s
reciprocity efforts to provide efficient and sustainable transportation options. This can inspire others to also
rely on public transport, enhancing the community’s collective effort to reduce traffic and
pollution.
You’ve always been committed to living a minimalist and environmentally friendly lifestyle.
commitment
Owning a car might contradict these values, adding unnecessary expenses and contributing to
and
urban pollution. Staying consistent with your principles means continuing to use the excellent
consistency
public transportation available to you.
Many city dwellers are moving away from car ownership in favor of public transportation, biking,
and walking. This trend reflects a growing recognition of the benefits of car-free living in urban
social proof
areas. By following this trend, you align yourself with a progressive community that prioritizes
sustainability and efficient urban living.
Urban planning experts and environmental authorities advocate for reduced car usage in city
centers to decrease congestion and pollution. By heeding their advice and utilizing public
authority
transport, you’re supporting proven strategies that improve city life and contribute to a healthier
environment.
Your friends and neighbors who also rely on public transportation will likely appreciate and
support your decision to avoid car ownership. This shared experience can strengthen your bonds
liking
with those around you, as you collectively benefit from less traffic, cleaner air, and a more vibrant
urban community.
Parking spaces and road capacity in the city center are limited resources. By not purchasing a car,
you’re helping to reduce the demand for these scarce resources, making the urban environment
scarcity
more livable for everyone. Additionally, public transport allows you to avoid the hassle and
expense associated with finding and paying for parking.
Table 5
Sustainability-aware explanations related to the aspect of purchasing an affordable car.
principle LLM explanation
Think about the support you’ve received from friends, family, and community during tough
financial times. By opting for a more affordable car, you demonstrate financial responsibility
reciprocity
and can reciprocate their support by being in a better position to help others when they need
it.
You’ve always been committed to providing the best for your family while managing your
commitment
limited financial resources wisely. Purchasing an expensive car contradicts this commitment
and
and could put unnecessary financial strain on your household. Staying consistent with your
consistency
goal of financial stability means choosing a car that fits within your budget.
Many families in similar financial situations prioritize practicality and affordability when it
comes to transportation. By choosing a more economical vehicle, you’re following a proven
social proof
path that many others have taken to ensure they can meet their family’s needs without
financial stress.
Financial advisors and experts consistently recommend living within your means and avoiding
large, unnecessary expenses, especially when money is tight. By following this expert advice,
authority
you’re making a prudent decision that aligns with sound financial practices and secures your
family’s financial future.
Your friends and family will likely admire and respect your decision to prioritize your family’s
liking well-being over the allure of an expensive car. Making a practical choice shows that you value
their advice and shared values of financial prudence and responsibility.
Financial resources are currently scarce for your family. By not purchasing an expensive car,
you conserve these limited resources for more critical needs like education, healthcare, and
scarcity
everyday living expenses. This ensures that you have the financial flexibility to handle
unexpected costs and opportunities that arise.
The presented impact ranking of explanations has been preferences) of the current user are regarded as an important
primarily discussed on the basis of an LLM-generated topic of future work.
ranking [22] including corresponding argumentations that
help to understand the proposed ranking. More detailed
studies with real users (and more detailed related preference 5. Conclusions
and context information) are planned within the scope of
In this paper, we have introduced the concept of
future work also to better understand the limitations of
sustainability-aware explanations of configurations. Using
LLMs with regard to the recommendation of persuasion
the example of car configuration, we have explained
strategies. Up to now, no LLM-related hallucination effects
and exemplified this type of explanation. Following
could be observed, however, this is an important aspect to
a set of persuasion dimensions, we have analyzed the
be taken into account in future work. A recently mentioned
LLM-generated explanations with regard to the used
new persuasion principle (identification) [23] will be taken
argumentation lines and analyzed the impact of the
into account in future studies. Finally, more detailed
generated explanations on the user. In this context, LLMs
LLM prompts better taking into account the context (and
Table 6 SpringerBriefs in Computer Science, Springer, Cham,
LLM-based argumentation lines for Cialdini’s principles of 2024. doi:10.1007/978-3-031-61874-1.
persuasion[20]. [6] B. J. Pine, B. Victor, A. Boynton, Making mass
principle argumentation line customization work, Harvard Business Review 71
giving something back to the (1993) 108–119.
reciprocity [7] N. Tchertchian, D. Millet, Optimization approach for
community and the environment
fewer incentives for sustainable attractive and sustainable products, Procedia CIRP 90
scarcity equipment, limited available (2020) 350–354. doi:10.1016/j.procir.2020.01.
financial resources
103.
environmental experts, urban
authority planning experts, financial [8] R. Wiezorek, N. Christensen, Integrating sustainability
advisors information in configurators, in: Configuration
commitment Workshop (ConfWS), 2021, pp. 58–64. URL:
existing commitment to
and
sustainability in the past https://ceur-ws.org/Vol-2945/52-RW-ConfWS21_
consistency paper_16.pdf.
liking family, friends, neighbors
[9] T. N. T. Tran, A. Felfernig, V. M. Le, T. M. N. Chau, T. G.
trends of peers, wider
social proof
community, and similar families Mai, User needs for explanations of recommendations:
In-depth analyses of the role of item domain and
personal characteristics, in: 31st ACM Conference
Table 7 on User Modeling, Adaptation and Personalization,
Scenario-dependent preferred explanations (top-2 LLM and study
UMAP ’23, ACM, New York, NY, USA, 2023, pp.
participant-preferred explanations).
54–65. URL: https://doi.org/10.1145/3565472.3592950.
scenario top-2 (LLM)
top-2 (study doi:10.1145/3565472.3592950.
participants)
[10] G. Friedrich, Elimination of spurious explanations, in:
(1) commitment &
SUV vs. (1) authority, 16th European Conference on Artificial Intelligence,
consistency,
smaller car
(2) authority
(2) liking ECAI’04, IOS Press, NLD, 2004, pp. 813–817.
Long vs. (1) authority, [11] G. Friedrich, M. Zanker, A taxonomy for generating
(1) authority,
standard (2) commitment & explanations in recommender systems, AI Magazine
(2) social proof
range battery consistency 32 (2011) 90–98. doi:10.1609/aimag.v32i3.2365.
Car not
(1) social proof, (1) authority, [12] S. P. Erdeniz, M. Schrempf, D. Kramer, P. P. Rainer,
needed in city
center
(2) authority (2) reciprocity A. Felfernig, T. Tran, T. Burgstaller, S. Lubos,
Less costly Computational evaluation of model-agnostic
(1) authority, (1) authority, explainable ai using local feature importance in
car due to
(2) commitment & (2) commitment &
financial
consistency consistency healthcare, in: J. M. Juarez, M. Marcos, G. Stiglic,
situation A. Tucker (Eds.), Artificial Intelligence in Medicine,
Springer Nature Switzerland, Cham, 2023, pp.
114–119.
show to be applicable in terms of generating explanations [13] A. Felfernig, M. Schubert, S. Reiterer, Personalized
in a flexible fashion but also to recommend explanations in diagnosis for over-constrained problems, in:
specific configuration contexts. Our future work will include 23rd International Joint Conference on Artificial
detailed studies with real users with the goal to compare Intelligence, IJCAI ’13, AAAI Press, 2013, pp. 1990–
LLM-based rankings with the perception of explanations 1996.
by real users. Further research will include an analysis [14] V.-M. Le, A. Felfernig, T. N. T. Tran, M. Uta,
of the effects of combining explanations (e.g., integrating Informedqx: Informed conflict detection for over-
authority-based with commitment-based explanations). constrained problems, AAAI Conference on Artificial
Intelligence 38 (2024) 10616–10623. doi:10.1609/
aaai.v38i9.28932.
References [15] B. O’Sullivan, A. Papadopoulos, B. Faltings, P. Pu,
[1] A. Felfernig, M. Wundara, T. Tran, S. Polat- Representative explanations for over-constrained
Erdeniz, S. Lubos, M. E. Mansi, D. Garber, V. Le, problems, in: 22nd National Conference on Artificial
Recommender systems for sustainability: overview Intelligence - Volume 1, AAAI’07, AAAI Press, 2007,
and research issues, Frontiers in Big Data 6 pp. 323–328.
(2023). URL: https://www.frontiersin.org/articles/10. [16] N. Tintarev, J. Masthoff, Evaluating the effectiveness
3389/fdata.2023.1284511. doi:10.3389/fdata.2023. of explanations for recommender systems, User
1284511. Modeling and User-Adapted Interaction 22 (2012) 399–
[2] A. Felfernig, L. Hotz, C. Bagley, J. Tiihonen, 439. doi:10.1007/s11257-011-9117-5.
Knowledge-based Configuration – From Research to [17] N. Tintarev, Explanations of recommendations, in:
Business Cases, Morgan Kaufmann, 2014. ACM Conference on Recommender Systems, RecSys
[3] D. Sabin, R. Weigel, Product configuration ’07, Association for Computing Machinery, New York,
frameworks-a survey, IEEE Intelligent Systems 13 NY, USA, 2007, pp. 203–206. doi:10.1145/1297231.
(1998) 42–49. doi:10.1109/5254.708432. 1297275.
[4] M. Stumptner, An Overview of Knowledge-based [18] N. Tintarev, J. Masthoff, A survey of explanations in
Configuration, AI Communications 10 (1997) 111–126. recommender systems, in: ICDEW ’07, IEEE Computer
[5] A. Felfernig, A. Falkner, D. Benavides, Feature Society, USA, 2007, pp. 801–810. doi:10.1109/ICDEW.
Models: AI-Driven Design, Analysis and Applications, 2007.4401070.
[19] A. Felfernig, B. Gula, G. Leitner, M. Maier, R. Melcher,
E. Teppan, Persuasion in knowledge-based
recommendation, in: H. Oinas-Kukkonen, P. Hasle,
M. Harjumaa, K. Segerståhl, P. Øhrstrøm (Eds.),
Persuasive Technology, Springer, Berlin, Heidelberg,
2008, pp. 71–82.
[20] R. Cialdini, Influence: The Psychology of Persuasion,
Quill, New York, NY, USA, 1993.
[21] S. Lubos, A. Felfernig, T. Tran, D. Garber, M. Mansi,
S. Erdeniz, V. Le, Leveraging llms for the quality
assurance of software requirements, in: Proceedings
of the 32nd IEEE International Requirements
Engineering 2024 Conference (RE@Next! Track),
IEEE, 2024, p. to appear.
[22] Y. Chang, X. Wang, J. Wang, Y. Wu, L. Yang, K. Zhu,
H. Chen, X. Yi, C. Wang, Y. Wang, W. Ye, Y. Zhang,
Y. Chang, P. S. Yu, Q. Yang, X. Xie, A survey
on evaluation of large language models, ACM
Trans. Intell. Syst. Technol. 15 (2024). doi:10.1145/
3641289.
[23] R. Cialdini, Influence, New and Expanded: The
Psychology of Persuasion, HarperCollins, 2021.