Publications | Αndreas Theodorou

2024

Pre Print
Governance

Navigating the sociotechnical labyrinth: Dynamic certification for responsible embodied AI

Georgios Bakirtzis, Andrea Aler Tubella, Andreas Theodorou, David Danks, and Ufuk Topcu

2024

arXiv
Journal

Governance

XAI

MLOps for Cyber-Physical Production Systems: Challenges and Solutions

L. Faubel, T. Woudsma, B. Kloepper, H. Eichelberger, F. Buelow, K. Schmid, and 4 more authors

IEEE Software, Aug 2024

Abs

Machine Learning Operations (MLOps) involves software development practices for Machine Learning (ML), including data management, preprocessing, model training, deployment, and monitoring. While MLOps have received significant interest, much less work has been published addressing MLOps in industrial production settings lately, particularly if solutions are not cloud-based. This article addresses this shortcoming based on our and our partner’s real industrial experience in various projects. While there is a broad range of challenges for MLOps in cyber-physical production systems (CPPS), we focus on those related to data, models, and operations as we assume these will directly benefit the reader and provide solutions such as lightweight integration, integration of domain knowledge, periodic calibration, and interactive interfaces. In this way, we want to support practitioners in setting up industrial MLOps environments in CPPS. Further, we discuss explainability as an additional part of MLOps, which should be explored in more detail in the future.
Journal

Variable Autonomy

Transparency

Who’s in Charge Here? A Survey on Trustworthy AI in Variable Autonomy Robotic Systems

Leila Methnani, Manolis Chiou, Virginia Dignum, and Andreas Theodorou

ACM Computing Surveys, Apr 2024

Abs

This article surveys the Variable Autonomy (VA) robotics literature that considers two contributory elements to Trustworthy AI: transparency and explainability. These elements should play a crucial role when designing and adopting robotic systems, especially in VA where poor or untimely adjustments of the system’s level of autonomy can lead to errors, control conflicts, user frustration, and ultimate disuse of the system. Despite this need, transparency and explainability is, to the best of our knowledge, mostly overlooked in VA robotics literature or is not considered explicitly. In this article, we aim to present and examine the most recent contributions to the VA literature concerning transparency and explainability. In addition, we propose a way of thinking about VA by breaking these two concepts down based on: the mission of the human-robot team; who the stakeholder is; what needs to be made transparent or explained; why they need it; and how it can be achieved. Last, we provide insights and propose ways to move VA research forward. Our goal with this article is to raise awareness and inter-community discussions among the Trustworthy AI and the VA robotics communities.
Proceedings

Argumentation
Clash of the explainers : argumentation for context-appropriate explanations

Leila Methnani, Virginia Dignum, and Andreas Theodorou

In Artificial Intelligence. ECAI 2023 : XAI^3, TACTIFUL, XI-ML, SEDAMI, RAAIT, AI4S, HYDRA, AI4AI, Kraków, Poland, September 30 - October 4, 2023, Proceedings, Part I , Apr 2024

Abs Bib

Understanding when and why to apply any given eXplainable Artificial Intelligence (XAI) technique is not a straightforward task. There is no single approach that is best suited for a given context. This paper aims to address the challenge of selecting the most appropriate explainer given the context in which an explanation is required. For AI explainability to be effective, explanations and how they are presented needs to be oriented towards the stakeholder receiving the explanation. If—in general—no single explanation technique surpasses the rest, then reasoning over the available methods is required in order to select one that is context-appropriate. Due to the transparency they afford, we propose employing argumentation techniques to reach an agreement over the most suitable explainers from a given set of possible explainers. In this paper, we propose a modular reasoning system consisting of a given mental model of the relevant stakeholder, a reasoner component that solves the argumentation problem generated by a multi-explainer component, and an AI model that is to be explained suitably to the stakeholder of interest. By formalizing supporting premises—and inferences—we can map stakeholder characteristics to those of explanation techniques. This allows us to reason over the techniques and prioritise the best one for the given context, while also offering transparency into the selection decision.
@inproceedings{MethnaniEtAl2024ClashOfExplainers, author = {Methnani, Leila and Dignum, Virginia and Theodorou, Andreas}, booktitle = {Artificial Intelligence. ECAI 2023 : XAI^3, TACTIFUL, XI-ML, SEDAMI, RAAIT, AI4S, HYDRA, AI4AI, Kraków, Poland, September 30 - October 4, 2023, Proceedings, Part I}, institution = {Umeå University, Department of Computing Science}, pages = {7--23}, title = {Clash of the explainers : argumentation for context-appropriate explanations}, series = {Communications in Computer and Information Science}, doi = {10.1007/978-3-031-50396-2_1}, isbn = {978-3-031-50396-2}, year = {2024}, keyword2 = {Argumentation} }

Book Chapter

Governance

Design

Responsible AI at work : incorporating human values

Andreas Theodorou, and Andrea Aler Tubella

In Handbook of artificial intelligence at work : interconnections and policy implications , Apr 2024

Bib

@incollection{TheodorouAler2024RAIAtWork,
  author = {Theodorou, Andreas and Aler Tubella, Andrea},
  booktitle = {Handbook of artificial intelligence at work : interconnections and policy implications},
  institution = {Umeå University, Department of Computing Science},
  pages = {32--46},
  title = {Responsible AI at work : incorporating human values},
  url = {https://www.e-elgar.com/shop/gbp/handbook-of-artificial-intelligence-at-work-9781800889965.html},
  isbn = {9781800889965},
  year = {2024},
  keyword1 = {Governance},
  keyword2 = {Design}
}

2023

Proceedings

Governance

Pedagogy
Operationalising AI ethics : conducting socio-technical assessment

Leila Methnani, Mattias Brännström, and Andreas Theodorou

In Human-Centered Artificial Intelligence : Advanced Lectures , Apr 2023

Part of the book sub series: Lecture Notes in Artificial Intelligence (LNAI)Conference series: ACAI: ECCAI Advanced Course on Artificial Intelligence

Abs Bib

Several high profile incidents that involve Artificial Intelligence (AI) have captured public attention and increased demand for regulation. Low public trust and attitudes towards AI reinforce the need for concrete policy around its development and use. However, current guidelines and standards rolled out by institutions globally are considered by many as high-level and open to interpretation, making them difficult to put into practice. This paper presents ongoing research in the field of Responsible AI and explores numerous methods of operationalising AI ethics. If AI is to be effectively regulated, it must not be considered as a technology alone—AI is embedded in the fabric of our societies and should thus be treated as a socio-technical system, requiring multi-stakeholder involvement and employment of continuous value-based methods of assessment. When putting guidelines and standards into practice, context is of critical importance. The methods and frameworks presented in this paper emphasise this need and pave the way towards operational AI ethics.
@inproceedings{MethnaniEtAl2023ACAI, author = {Methnani, Leila and Br{\"a}nnstr{\"o}m, Mattias and Theodorou, Andreas}, booktitle = {Human-Centered Artificial Intelligence : Advanced Lectures}, institution = {Umeå University, Department of Computing Science}, note = {Part of the book sub series: Lecture Notes in Artificial Intelligence (LNAI)Conference series: ACAI: ECCAI Advanced Course on Artificial Intelligence}, pages = {304--321}, title = {Operationalising AI ethics : conducting socio-technical assessment}, series = {Lecture Notes in Computer Science}, number = {13500}, doi = {10.1007/978-3-031-24349-3_16}, isbn = {9783031243486}, year = {2023}, keyword1 = {Governance}, keyword2 = {Pedagogy} }
Proceedings

Variable Autonomy
Variable Autonomy for Human-Robot Teaming (VAT)

Manolis Chiou, Serena Booth, Bruno Lacerda, Andreas Theodorou, and Simon Rothfuß

In HRI ’23 : Companion of the 2023 ACM/IEEE International Conference on Human-Robot Interaction , Apr 2023

Abs Bib

As robots are introduced to various domains and applications, Human-Robot Teaming (HRT) capabilities are essential. Such capabilities involve teaming with humans in/on/out-the-loop at different levels of abstraction, leveraging the complementing capabilities of humans and robots. This requires robotic systems with the ability to dynamically vary their level or degree of autonomy to collaborate with the human(s) efficiently and overcome various challenging circumstances. Variable Autonomy (VA) is an umbrella term encompassing such research, including but not limited to shared control and shared autonomy, mixed-initiative, adjustable autonomy, and sliding autonomy. This workshop is driven by the timely need to bring together VA-related research and practices that are often disconnected across different communities as the field is relatively young. The workshop’s goal is to consolidate research in VA. To this end, and given the complexity and span of Human-Robot systems, this workshop will adopt a holistic trans-disciplinary approach aiming to a) identify and classify related common challenges and opportunities; b) identify the disciplines that need to come together to tackle the challenges; c) identify and define common terminology, approaches, methodologies, benchmarks, and metrics; d) define short- and longterm research goals for the community. To achieve these objectives, this workshop aims to bring together industry stakeholders, researchers from fields under the banner of VA, and specialists from other highly related fields such as human factors and psychology. The workshop will consist of a mix of invited talks, contributed papers, and an interactive discussion panel, toward a shared vision for VA.
@inproceedings{ChiouEtAl2023VAT, author = {Chiou, Manolis and Booth, Serena and Lacerda, Bruno and Theodorou, Andreas and Rothfuß, Simon}, booktitle = {HRI '23 : Companion of the 2023 ACM/IEEE International Conference on Human-Robot Interaction}, institution = {Karlsruhe Institute of Technology, Karlsruhe, Germany}, pages = {932--932}, title = {Variable Autonomy for Human-Robot Teaming (VAT)}, series = {ACM/IEEE International Conference on Human-Robot Interaction}, doi = {10.1145/3568294.3579957}, isbn = {978-1-4503-9970-8}, year = {2023}, keyword1 = {Variable Autonomy} }

Editorial

Governance

From fear to action : AI governance and opportunities for all

Kevin Baum , Joanna Bryson, Frank Dignum, Virginia Dignum, Marko Grobelnik, Holger Hoos, and 7 more authors

Frontiers in Computer Science, Apr 2023

Bib

@article{BaumEtAl2023FromFearToAction,
  author = {Baum, Kevin and Bryson, Joanna and Dignum, Frank and Dignum, Virginia and Grobelnik, Marko and Hoos, Holger and Irgens, Morten and Lukowicz, Paul and Muller, Catelijne and Rossi, Francesca and Shawe-Taylor, John and Theodorou, Andreas and Vinuesa, Ricardo},
  journal = {Frontiers in Computer Science},
  eid = {1210421},
  title = {From fear to action : AI governance and opportunities for all},
  volume = {5},
  doi = {10.3389/fcomp.2023.1210421},
  year = {2023},
  keyword1 = {Governance}
}

Journal

Governance
Garbage in, toxic data out : a proposal for ethical artificial intelligence sustainability impact statements

Ronny Bogani, Andreas Theodorou, Luca Arnaboldi , and Robert H. Wortham

AI and Ethics, Apr 2023

Published online: 20 October 2022

Abs Bib

Data and autonomous systems are taking over our lives, from healthcare to smart homes very few aspects of our day to day are not permeated by them. The technological advances enabled by these technologies are limitless. However, with advantages so too come challenges. As these technologies encompass more and more aspects of our lives, we are forgetting the ethical, legal, safety and moral concerns that arise as an outcome of integrating our lives with technology. In this work, we study the lifecycle of artificial intelligence from data gathering to deployment, providing a structured analytical assessment of the potential ethical, safety and legal concerns. The paper then presents the foundations for the first ethical artificial intelligence sustainability statement to guide future development of AI in a safe and sustainable manner.
@article{Bogani1706385, author = {Bogani, Ronny and Theodorou, Andreas and Arnaboldi, Luca and Wortham, Robert H.}, institution = {Department of Electrical Engineering, University of Bath, Bath, UK}, journal = {AI and Ethics}, note = {Published online: 20 October 2022}, pages = {1135--1142}, title = {Garbage in, toxic data out : a proposal for ethical artificial intelligence sustainability impact statements}, volume = {3}, doi = {10.1007/s43681-022-00221-0}, year = {2023}, keyword1 = {Governance} }

2022

Proceedings

Governance

Audit

Let it RAIN for Social Good

Mattias Brännström, Andreas Theodorou, and Virginia Dignum

In IJCAI 2022 Workshop on AISafety , Apr 2022

arXiv Bib

@inproceedings{BrannstromEtAl2022AISafety,
  author = {Brännström, Mattias and Theodorou, Andreas and Dignum, Virginia},
  title = {Let it {RAIN} for Social Good},
  booktitle = {IJCAI 2022 Workshop on AISafety},
  year = {2022},
  archiveprefix = {arXiv},
  keyword1 = {Governance},
  keyword2 = {Audit}
}

Proceedings

Games AI

MAS

Embracing AWKWARD! Real-time Adjustment of Reactive Planning Using Social Norms

Leila Methnani, Andreas Antoniades, and Andreas Theodorou

In Proceedings of the Coordination, Organizations, Institutions, Norms, and Ethics for Governance of Multi-Agent Systems (COINE) 2022 , Apr 2022

Bib

@inproceedings{MethnaniEtAl2022COINE,
  author = {Methnani, Leila and Antoniades, Andreas and Theodorou, Andreas},
  title = {Embracing {AWKWARD}! Real-time Adjustment of Reactive Planning Using Social Norms},
  booktitle = {Proceedings of the Coordination, Organizations, Institutions, Norms, and Ethics for Governance of Multi-Agent Systems (COINE) 2022},
  year = {2022},
  keyword1 = {Games AI},
  keyword2 = {MAS},
  keyword3 = {Cogn. Arch.}
}

Journal

Governance

Fairness

A sociotechnical perspective for the future of AI: narratives, inequalities, and human control

Laura Sartori, and Andreas Theodorou

Ethics of Information Technology, Apr 2022

Bib

@article{SartoriTheodorou2022,
  author = {Sartori, Laura and Theodorou, Andreas},
  title = {A sociotechnical perspective for the future of AI: narratives, inequalities, and human control},
  journal = {Ethics of Information Technology},
  year = {2022},
  doi = {10.1007/s10676-022-09624-3},
  url = {https://link.springer.com/article/10.1007/s10676-022-09624-3},
  volume = {24},
  keyword1 = {Governance},
  keyword2 = {Fairness},
  keyword3 = {Transparency}
}

2021

Journal

Variable Autonomy

Governance

Let Me Take Over: Variable Autonomy For Meaningful Human Control

Leila Methnani, Andrea Aler Tubella, Virginia Dignum, and Andreas Theodorou

Frontiers in Artificial Intelligence, Apr 2021

Bib

@article{Methnani2021Frontiers,
  author = {Methnani, Leila and {Aler Tubella}, Andrea and Dignum, Virginia and Theodorou, Andreas},
  title = {Let Me Take Over: Variable Autonomy For Meaningful Human Control},
  journal = {Frontiers in Artificial Intelligence},
  year = {2021},
  volume = {4},
  pages = {133},
  url = {https://www.frontiersin.org/article/10.3389/frai.2021.737072},
  doi = {10.3389/frai.2021.737072},
  issn = {2624-8212},
  keyword1 = {Variable Autonomy},
  keyword2 = {Governance}
}

Journal

Governance

Transparency
IEEE P7001: A Proposed Standard on Transparency

Alan F. T. Winfield, Serena Booth, Louise A. Dennis, Takashi Egawa, Helen Hastie, Naomi Jacobs, and 7 more authors

Frontiers in Robotics and AI, Apr 2021

Abs Bib

This paper describes IEEE P7001, a new draft standard on transparency of autonomous systems¹. In the paper, we outline the development and structure of the draft standard. We present the rationale for transparency as a measurable, testable property. We outline five stakeholder groups: users, the general public and bystanders, safety certification agencies, incident/accident investigators and lawyers/expert witnesses, and explain the thinking behind the normative definitions of “levels” of transparency for each stakeholder group in P7001. The paper illustrates the application of P7001 through worked examples of both specification and assessment of fictional autonomous systems.
@article{WinfieldEtAl2021Frontiers, author = {Winfield, {Alan F. T}. and Booth, {Serena} and Dennis, {Louise A.} and Egawa, Takashi and Hastie, Helen and Jacobs, Naomi and Muttram, {Roderick I.} and Olszewska, {Joanna I.} and Rajabiyazdi, Fahimeh and Theodorou, Andreas and Underwood, {Mark A.} and Wortham, {Robert H.} and Watson, Eleanor}, title = {IEEE P7001: A Proposed Standard on Transparency}, journal = {Frontiers in Robotics and AI}, volume = {8}, issue = {665729}, year = {2021}, url = {https://www.frontiersin.org/article/10.3389/frobt.2021.665729}, doi = {10.3389/frobt.2021.665729}, issn = {2296-9144}, keyword1 = {Governance}, keyword2 = {Transparency} }
Proceedings

Governance

Transparency

Interrogating the Black Box: Transparency through Information-Seeking Dialogues

Andrea Aler Tubella, Andreas Theodorou, and Juan Carlos Nieves

In Proceedings of the 20th International Conference on Autonomous Agents and MultiAgent Systems , Apr 2021

Abs arXiv

This paper is preoccupied with the following question: given a (possibly opaque) learning system, how can we understand whether its behaviour adheres to governance constraints? The answer can be quite simple: we just need to "ask" the system about it. We propose to construct an investigator agent to query a learning agent– the suspect agent– to investigate its adherence to a given ethical policy in the context of an information-seeking dialogue, modeled in formal argumentation settings. This formal dialogue framework is the main contribution of this paper. Through it, we break down compliance checking mechanisms into three modular components, each of which can be tailored to various needs in a vast amount of ways: an investigator agent, a suspect agent, and an acceptance protocol determining whether the responses of the suspect agent comply with the policy. This acceptance protocol presents a fundamentally different approach to aggregation: rather than using quantitative methods to deal with the non-determinism of a learning system, we leverage the use of argumentation semantics to investigate the notion of properties holding consistently. Overall, we argue that the introduced formal dialogue framework opens many avenues both in the area of compliance checking and in the analysis of properties of opaque systems.

2020

Report
Governance

Final Anaysis on the EU Whitepaper on AI

Catelijne Muller, Virginia Dignum, and Andreas Theodorou

Apr 2020

Proceedings

Governance

Audit

Contestable Black Boxes

Andrea Aler Tubella, Andreas Theodorou, Virginia Dignum, and Loizos Michael

In 4th International Joint Conference on Rules and Reasoning (RuleML) , Jun 2020

arXiv Bib

@inproceedings{AlerTubellaEtAl2020RULEML,
  author = {{Aler Tubella}, Andrea and Theodorou, Andreas and Dignum, Virginia and Michael, Loizos},
  booktitle = {4th International Joint Conference on Rules and Reasoning (RuleML)},
  month = jun,
  address = {Olso, Norway},
  title = {{Contestable Black Boxes}},
  doi = {10.1007/978-3-030-57977-7_12},
  year = {2020},
  keyword1 = {Governance},
  keyword2 = {Audit}
}

Journal

Governance

Audit

A socio-technical framework for digital contact tracing

Ricardo Vinuesa, Andreas Theodorou, Manuela Battaglini, and Virginia Dignum

Results in Engineering, Jun 2020

arXiv Bib

@article{VinuesaEtAl2020SocioTechnical,
  title = {A socio-technical framework for digital contact tracing},
  author = {Vinuesa, Ricardo and Theodorou, Andreas and Battaglini, Manuela and Dignum, Virginia},
  year = {2020},
  journal = {Results in Engineering},
  doi = {10.1016/j.rineng.2020.100163},
  keyword1 = {Governance},
  keyword2 = {Audit}
}

Journal

Governance

Towards ethical and socio-legal governance in AI

Andreas Theodorou, and Virginia Dignum

Nature Machine Intelligence, Jun 2020

Bib

@article{TheodorouDignum2020NMI,
  author = {Theodorou, Andreas and Dignum, Virginia},
  year = {2020},
  doi = {10.1038/s42256-019-0136-y6},
  journal = {Nature Machine Intelligence},
  title = {{Towards ethical and socio-legal governance in AI}},
  volume = {2},
  issue = {1},
  url = {{https://www.nature.com/articles/s42256-019-0136-y.epdf?shared_access_token=tLQjOyEeJJBLNYBpYAddZ9RgN0jAjWel9jnR3ZoTv0OKielulWtfIKdoTkc7o23A4ag4RzhLocCFIkMqRYeFumYGAnLqPSfK_tQ3761isKFC32POZ17DGXFsQMNDEcD8X2AnDXspfKQtpudOtnxcvQ%3D%3D}},
  keyword1 = {Governance}
}

Book Chapter

Governance

Why Artificial Intelligence is a matter of Design

Andreas Theodorou

In Artificial Intelligence: Reflections in Philosophy, Theology, and Social Sciences , Jun 2020

Bib

@incollection{Theodorou2020AIReflections,
  author = {Theodorou, Andreas},
  booktitle = {Artificial Intelligence: Reflections in Philosophy, Theology, and Social Sciences},
  editor = {Goecke, Benedikt Paul and Rosenthal-von der P{\"{u}}tten, Astrid Marieke},
  publisher = {Brill},
  title = {{Why Artificial Intelligence is a matter of Design}},
  doi = {10.30965/9783957437488_009},
  year = {2020},
  keyword1 = {Governance}
}

2019

Book Chapter

Governance

Cogn. Arch.
How Society Can Maintain Human-Centric Artificial Intelligence

Joanna J Bryson, and Andreas Theodorou

In Human-Centered Digitalization and Services , Jun 2019

Abs Bib

Although not a universally held goal, maintaining human-centric artificial intelligence is necessary for society’s long-term stability. Fortunately, the legal and technological problems of maintaining control are actually fairly well understood and amenable to engineering. The real problem is establishing the social and political will for assigning and maintaining accountability for artifacts when these artefacts are generated or used. In this chapter we review the necessity and tractability of maintaining human control, and the mechanisms by which this can be achieved. What makes the problem both most interesting and most threatening is that achieving consensus around such an approach requires at least some measure of agreement on broad existential concerns.
@incollection{BrysonTheodorou2019, author = {Bryson, Joanna J and Theodorou, Andreas}, booktitle = {Human-Centered Digitalization and Services}, doi = {10.1007/978-981-13-7725-9_16}, editor = {Toivonen-Noro, Marja and Saari, Evelina and Melkas, Helin{\"{a}} and Hasu, Mervin}, pages = {305--323}, publisher = {Springer}, title = {{How Society Can Maintain Human-Centric Artificial Intelligence}}, url = {http://www.cs.bath.ac.uk/{~}jjb/web/publications.html http://link.springer.com/10.1007/978-981-13-7725-9{\_}16}, year = {2019}, keyword1 = {Governance}, keyword2 = {Cogn. Arch.} }

Proceedings

HRI

Transparency

Improving Robot Transparency: An Investigation With Mobile Augmented Reality

Alexandros Rotsidis, Andreas Theodorou , Joanna J. Bryson , and Robert H. Wortham

In 2019 28th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) , Oct 2019

Bib

@inproceedings{RotsidisEtAl2019ROMAN,
  address = {New Delhi, India},
  author = {Rotsidis, Alexandros and Theodorou, Andreas and Bryson, Joanna J. and Wortham, Robert H.},
  booktitle = {2019 28th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN)},
  doi = {10.1109/RO-MAN46459.2019.8956390},
  isbn = {978-1-7281-2622-7},
  month = oct,
  pages = {1--8},
  publisher = {IEEE},
  title = {{Improving Robot Transparency: An Investigation With Mobile Augmented Reality}},
  url = {https://ieeexplore.ieee.org/document/8956390/},
  year = {2019},
  keyword1 = {HRI},
  keyword2 = {Transparency}
}

Thesis
AI Governance Through a Transparency Lens

Andreas Theodorou

University of Bath , Oct 2019

Abs Bib

Transparency is a key consideration for the ethical design and use of Artificial Intelligence, and has recently become a topic of considerable public interest and debate. We frequently use philosophical, mathematical, and biologically inspired techniques for building artificial, interactive, intelligent agents. Yet despite these well-motivated inspirations, the resulting intelligence is often developed as a black box, communicating no understanding of how the underlying real-time decision making functions. This compromises both the safety of such systems and fair attribution of moral responsibility and legal accountability when incidents occur. This dissertation provides the knowledge and software tools to make artificially intelligent agents more transparent, allowing a direct understanding of the action-selection system of such a system. The use of transparency, as demonstrated in this document, helps not only with the debugging of intelligent agents, but also with the public’s understanding of Artificial Intelligence (AI) by removing the ’scary’ mystery around "why is it behaving like that". In the research described in this document I investigate and compare the perception we have of intelligent systems, such as robots and autonomous vehicles, when they are treated as black boxes compared to when we make their action-selection systems transparent. Finally, I make normative and descriptive arguments for the moral status of intelligent systems and contribute to regulatory policy regarding such systems.
@phdthesis{Theodorou2019Thesis, author = {Theodorou, Andreas}, url = {{https://github.com/RecklessCoding/recklesscoding.github.io/raw/master/files/andreasTheodorouThesis.pdf}}, school = {University of Bath}, title = {{AI Governance Through a Transparency Lens}}, year = {2019}, }

Proceedings

Games ABM

The Sustainability Game : AI Technology as an Intervention for Public Understanding of Cooperative Investment

Andreas Theodorou, Bryn Bandt-law , and Joanna J Bryson

In IEEE Conference on Games , Oct 2019

Bib

@inproceedings{Theodorou2019COG,
  author = {Theodorou, Andreas and Bandt-law, Bryn and Bryson, Joanna J},
  booktitle = {IEEE Conference on Games},
  title = {{The Sustainability Game : AI Technology as an Intervention for Public Understanding of Cooperative Investment}},
  url = {https://bibbase.org/service/mendeley/3de73d64-9180-31af-9719-79efb2b3294b/file/7295b6eb-5102-39b9-0730-fdd46d5c2553/TheodorouEtAl{\_}TheSustainabilityGame{\_}COG2019{\_}SemiFinal.pdf.pdf},
  year = {2019},
  keyword1 = {Games},
  keyword2 = {ABM}
}

Proceedings

Governance

Audit
Governance by Glass-Box: Implementing Transparent Moral Bounds for AI Behaviour

Andrea Aler Tubella, Andreas Theodorou, Virginia Dignum, and Frank Dignum

In International Joint Conference on Artificial Intellignece (IJCAI) , Oct 2019

Abs arXiv Bib

Artificial Intelligence (AI) applications are being used to predict and assess behaviour in multiple domains, such as criminal justice and consumer finance, which directly affect human well-being. However, if AI is to improve people’s lives, then people must be able to trust AI, which means being able to understand what the system is doing and why. Even though transparency is often seen as the requirement in this case, realistically it might not always be possible or desirable, whereas the need to ensure that the system operates within set moral bounds remains. In this paper, we present an approach to evaluate the moral bounds of an AI system based on the monitoring of its inputs and outputs. We place a "glass box" around the system by mapping moral values into explicit verifiable norms that constrain inputs and outputs, in such a way that if these remain within the box we can guarantee that the system adheres to the value. The focus on inputs and outputs allows for the verification and comparison of vastly different intelligent systems; from deep neural networks to agent-based systems. The explicit transformation of abstract moral values into concrete norms brings great benefits in terms of explainability; stakeholders know exactly how the system is interpreting and employing relevant abstract moral human values and calibrate their trust accordingly. Moreover, by operating at a higher level we can check the compliance of the system with different interpretations of the same value. These advantages will have an impact on the well-being of AI systems users at large, building their trust and providing them with concrete knowledge on how systems adhere to moral values.
@inproceedings{AlerTubellaEtAl2019IJCAI, archiveprefix = {arXiv}, author = {{Aler Tubella}, Andrea and Theodorou, Andreas and Dignum, Virginia and Dignum, Frank}, booktitle = {International Joint Conference on Artificial Intellignece (IJCAI)}, eprint = {1905.04994}, title = {{Governance by Glass-Box: Implementing Transparent Moral Bounds for AI Behaviour}}, url = {http://arxiv.org/abs/1905.04994}, year = {2019}, keyword1 = {Governance}, keyword2 = {Audit} }

Proceedings

HRI

Transparency

Slam the Brakes: Perceptions of Moral Decisions in Driving Dilemmas

Holly Wilson, and Andreas Theodorou

In IJCAI 2019 Workshop on AISafety , Oct 2019

Bib

@inproceedings{WilsonTheodorou2019,
  address = {Macao, China},
  author = {Wilson, Holly and Theodorou, Andreas},
  booktitle = {IJCAI 2019 Workshop on AISafety},
  title = {{Slam the Brakes: Perceptions of Moral Decisions in Driving Dilemmas}},
  url = {{http://ceur-ws.org/Vol-2419/paper_14.pdf}},
  year = {2019},
  keyword1 = {HRI},
  keyword2 = {Transparency},
  keyword3 = {XAI}
}

Proceedings

HRI

Transparency
Robots that make sense: Transparent intelligence through augmented reality

Alexandros Rotsidis, Andreas Theodorou , and Robert H. Wortham

In IUI’19 Workshop on Intelligent User Interfaces for Algorithmic Transparency in Emerging Technologies , Oct 2019

Abs Bib

Autonomous robots can be difficult to understand by their develop-ers, let alone by end users. Yet, as they become increasingly integralparts of our societies, the need for affordable easy to use tools toprovide transparency grows. The rise of the smartphone and theimprovements in mobile computing performance have graduallyallowed Augmented Reality (AR) to become more mobile and afford-able. In this paper we review relevant robot systems architectureand propose a new software tool to provide robot transparencythrough the use of AR technology. Our new tool, ABOD3-AR pro-vides real-time graphical visualisation and debugging of a robot’sgoals and priorities as a means for both designers and end usersto gain a better mental model of the internal state and decisionmaking processes taking place within a robot. We also report onour on-going research programme and planned studies to furtherunderstand the effects of transparency to naive users and experts.
@inproceedings{RotsidisEtAl2019IUI, title = {Robots that make sense: Transparent intelligence through augmented reality}, year = {2019}, volume = {2327}, address = {Los Angeles, CA USA}, author = {Rotsidis, Alexandros and Theodorou, Andreas and Wortham, Robert H.}, booktitle = {IUI'19 Workshop on Intelligent User Interfaces for Algorithmic Transparency in Emerging Technologies}, keyword1 = {HRI}, keyword2 = {Transparency} }

2017

Journal

Governance

Transparency
Designing and implementing transparency for real time inspection of autonomous robots

Andreas Theodorou , Robert H. Wortham , and Joanna J. Bryson

Connection Science, Oct 2017

Abs Bib

The EPSRC’s Principles of Robotics advises the implementation of transparency in robotic systems, however research related to AI transparency is in its infancy. This paper introduces the reader of the importance of having transparent inspection of intelligent agents and provides guidance for good practice when developing such agents. By considering and expanding upon other prominent definitions found in literature, we provide a robust definition of transparency as a mechanism to expose the decision-making of a robot. The paper continues by addressing potential design decisions developers need to consider when designing and developing transparent systems. Finally, we describe our new interactive intelligence editor, designed to visualise, develop and debug real-time intelligence.
@article{Theodorou2017ConnectionScience, author = {Theodorou, Andreas and Wortham, Robert H. and Bryson, Joanna J.}, doi = {10.1080/09540091.2017.1310182}, issn = {13600494}, journal = {Connection Science}, number = {3}, pages = {230--241}, title = {{Designing and implementing transparency for real time inspection of autonomous robots}}, url = {https://researchportal.bath.ac.uk/files/154473870/TheodorouDesigningAndImplementingTransparency.pdf}, volume = {29}, year = {2017}, keyword1 = {Governance}, keyword2 = {Transparency} }
Journal

Governance

Transparency
Robot transparency, trust and utility

Robert H. Wortham, and Andreas Theodorou

Connection Science, Oct 2017

Abs Bib

As robot reasoning becomes more complex, debugging becomes increasingly hard based solely on observable behaviour, even for robot designers and technical specialists. Similarly, non-specialist users find it hard to create useful mental models of robot reasoning solely from observed behaviour. The EPSRC Principles of Robotics mandate that our artefacts should be transparent, but what does this mean in practice, and how does transparency affect both trust and utility? We investigate this relationship in the literature and find it to be complex, particularly in non industrial environments where transparency may have a wider range of effects on trust and utility depending on the application and purpose of the robot. We outline our programme of research to support our assertion that it is nevertheless possible to create transparent agents that are emotion-ally engaging despite having a transparent machine nature.
@article{Wortham2017ConnScience, author = {Wortham, Robert H. and Theodorou, Andreas}, doi = {10.1080/09540091.2017.1313816}, issn = {13600494}, journal = {Connection Science}, number = {3}, pages = {242--248}, title = {{Robot transparency, trust and utility}}, url = {https://researchportal.bath.ac.uk/files/154037267/transparency.pdf}, volume = {29}, year = {2017}, keyword1 = {Governance}, keyword2 = {Transparency} }

Proceedings

HRI

Transparency

Improving robot transparency: Real-Time visualisation of robot AI substantially improves understanding in naive observers

Robert H. Wortham, Andreas Theodorou , and Joanna J. Bryson

In RO-MAN 2017 - 26th IEEE International Symposium on Robot and Human Interactive Communication , Oct 2017

Bib

@inproceedings{Wortham2017ROMAN,
  author = {Wortham, Robert H. and Theodorou, Andreas and Bryson, Joanna J.},
  booktitle = {RO-MAN 2017 - 26th IEEE International Symposium on Robot and Human Interactive Communication},
  doi = {10.1109/ROMAN.2017.8172491},
  isbn = {9781538635186},
  issn = {1528-1132 0009-921X},
  pages = {1424--1431},
  publisher = {IEEE},
  title = {{Improving robot transparency: Real-Time visualisation of robot AI substantially improves understanding in naive observers}},
  url = {https://researchportal.bath.ac.uk/files/155724003/robot{\_}transparency{\_}experiment.pdf},
  volume = {2017-Janua},
  year = {2017},
  keyword1 = {HRI},
  keyword2 = {Transparency}
}

Proceedings

HRI

Transparency
Robot transparency: Improving understanding of intelligent behaviour for designers and users

Robert H. Wortham, Andreas Theodorou , and Joanna J. Bryson

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Oct 2017

Abs Bib

Autonomous robots can be difficult to design and understand. Designers have difficulty decoding the behaviour of their own robots simply by observing them. Naive users of robots similarly have diculty decoding robot behaviour simply through observation. In this paper we review relevant robot systems architecture, design and transparency literature, and report on a programme of research to investigate practical approaches to improve robot transparency. We report on the investigation of real-time graphical and vocalised outputs as a means for both designers and end users to gain a better mental model of the internal state and decision making processes taking place within a robot. This approach, combined with a graphical approach to behaviour design, o ers improved transparency for robot designers. We also report on studies of users’ understanding, where significant improvement has been achieved using both graphical and vocalisation transparency approaches.
@article{WorthamEtAl2017TAROS, author = {Wortham, Robert H. and Theodorou, Andreas and Bryson, Joanna J.}, doi = {10.1007/978-3-319-64107-2_22}, isbn = {9783319641065}, issn = {16113349}, journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}, pages = {274--289}, title = {{Robot transparency: Improving understanding of intelligent behaviour for designers and users}}, volume = {10454}, year = {2017}, keyword1 = {HRI}, keyword2 = {Transparency} }

2016

Proceedings

Governance

Transparency

ABOD3: A graphical visualisation and real-time debugging tool for bod agents

Andreas Theodorou

In CEUR Workshop Proceedings , Oct 2016

Bib

@inproceedings{Theodorou2017EUCog,
  author = {Theodorou, Andreas},
  booktitle = {CEUR Workshop Proceedings},
  issn = {16130073},
  pages = {60--61},
  title = {{ABOD3: A graphical visualisation and real-time debugging tool for bod agents}},
  volume = {1855},
  year = {2016},
  keyword1 = {Governance},
  keyword2 = {Transparency}
}

Proceedings

HRI

Transparency
What Does the Robot Think? Transparency as a Fundamental Design Requirement for Intelligent Systems

Robert H. Wortham, Andreas Theodorou , and Joanna J. Bryson

In IJCAI 2016 Ethics for Artificial Intelligence Workshop , Oct 2016

Abs Bib

Deciphering the behaviour of intelligent others is a fundamental characteristic of our own intelligence. As we interact with complex intelligent artefacts, humans inevitably construct mental models to understand and predict their behaviour. If these models are incorrect or inadequate, we run the risk of self deception or even harm. This paper reports progress on a programme of work investigating approaches for implementing robot transparency, and the effects of these approaches on utility, trust and the perception of agency. Preliminary findings indicate that building transparency into robot action selection can help users build a more accurate understanding of the robot.
@inproceedings{WorthamEtAl2016IJCAI, author = {Wortham, Robert H. and Theodorou, Andreas and Bryson, Joanna J.}, booktitle = {IJCAI 2016 Ethics for Artificial Intelligence Workshop}, title = {{What Does the Robot Think? Transparency as a Fundamental Design Requirement for Intelligent Systems}}, url = {https://researchportal.bath.ac.uk/en/publications/what-does-the-robot-think-transparency-as-a-fundamental-design-re}, year = {2016}, address = {New York, NY, US}, keyword1 = {HRI}, keyword2 = {Transparency} }