Existing AI ethics frameworks produce principles rather than instruments. They describe what responsible AI should look like without providing development teams with a procedure for measuring whether their system is getting there. This paper presents the Ethics Readiness Levels (ERL) framework, an evaluation methodology that adapts the Technology Readiness Level paradigm to the domain of AI ethics. ERL defines a maturity scale (ERL 0–4) derived from Integration Readiness Levels, implements it through an indicator questionnaire with tree-structured navigation, and tracks progress across development milestones. The methodology operationalizes thick ethical concepts into indicators. It uses a risk-mitigation-validation scoring mechanism that rewards validated mitigation. It requires dialogue between an ethics expert and a technical expert to encourage ethical reflection and correct for self-evaluation bias. The framework is designed to function as a management tool for ethics maturity. We describe the framework’s architecture, its conceptual foundations, and its strengths and limitations.

Quantum technologies (QT) are advancing rapidly, promising advancements across a wide spectrum of applications but also raising significant ethical, societal, and geopolitical impacts, including dual-use capabilities, varying levels of access, and impending quantum divide(s). To address these, the Responsible Quantum Technologies (ResQT) community was established to share knowledge, perspectives, and best practices across various disciplines. Its mission is to ensure QT developments align with ethical principles, promote equity, and mitigate unintended consequences. Initial progress has been made, as scholars and policymakers increasingly recognize principles of responsible QT. However, more widespread dissemination is needed, and as QT matures, so must responsible QT. This paper provides a structured, situated synthesis of the ResQT community's current work and identifies future directions from within this community perspective. Drawing on historical lessons from artificial intelligence and nanotechnology, actions targeting the quantum divide(s) are addressed, including the implementation of responsible research and innovation, fostering wider stakeholder engagement, and sustainable development. These actions aim to build trust and engagement, facilitating the participatory and responsible development of QT. The ResQT community advocates that responsible QT should be an integral part of quantum development rather than an afterthought so that quantum technologies evolve toward a future that is technologically advanced and beneficial for all.

We present Ethics Readiness Levels (ERLs), a four-level, iterative method to track how ethical reflection is implemented in the design of AI systems. ERLs bridge high-level ethical principles and everyday engineering by turning ethical values into concrete prompts, checks, and controls within real use cases. The evaluation is conducted using a dynamic, tree-like questionnaire built from context-specific indicators, ensuring relevance to the technology and application domain. Beyond being a managerial tool, ERLs help facilitate a structured dialogue between ethics experts and technical teams, while our scoring system helps track progress over time. We demonstrate the methodology through two case studies: an AI facial sketch generator for law enforcement and a collaborative industrial robot. The ERL tool effectively catalyzes concrete design changes and promotes a shift from narrow technological solutionism to a more reflective, ethics-by-design mindset.

After two decades of research on indefinite causality, a philosophical lesson emerges: the tension between operational quantum theory and dynamical spacetime physics is unbridgeable if one believes both types of theories to be fundamental. We interpret this tension through operational eternalism, a stance analogous to the block-universe view but applied to information rather than geometry. Inputs and outputs are primary givens, while agents are secondary constructs arising from specific groupings of data. Agency is perspectival: from Alice's perspective Bob may not qualify as an observer, and vice versa. These results redefine the observer in the operational approach as a tool to avoid non-causality. They also provide a criterion for Wigner's friends as a class of causally compatible agents.

We present a training module in AI ethics designed to prepare a broad group of professionals to recognize and address potential ethical challenges of AI applications in healthcare. Training materials include a two-page checklist, a brief glossary, and three practical case studies. While we have developed and applied this framework for training Research Ethics Committee members in France and South Africa, it can also be helpful in university courses ranging from public health and healthcare law to biomedical engineering and applied ethics.

This article delves into the implications of neurotechnologies for the philosophical debates surrounding free will and moral responsibility. Tracing the concept from ancient religious and philosophical roots, we discuss how recent neurotechnological advancements (e.g. optogenetics, fMRI and machine learning, predictive diagnostics, et al.) challenge traditional notions of autonomy. Although neurotechnologies aim to enhance autonomy in the strict sense – as self-determination – they risk reducing or changing the broader notion of autonomy, which involves personal authenticity. We also submit that, in a world with an altered or limited concept of free will, humans should still be held accountable for actions executed through their bodies. By examining the dynamic between choice and responsibility, we emphasize the shift in technology ethics, moral philosophy, and the broader legal landscape in response to the advancement of neurotechnologies. By bringing the neurotechnological innovations into the world, neuroscientists not only change the technological landscape but also partake in long-standing moral narratives about freedom, justice, and responsibility.

We propose two new readiness level scales. The first scale is dedicated to security aspects that accompany technological development. It follows the cybersecurity capability model. The second scale indicates the legal, privacy, and ethics readiness of technological components based on integrating human-centric considerations into a design. The scale derives its formal features from the integration readiness level structure but seeks to encompass a broader ethics-by-design approach. We argue that both security and ethics readiness scales deserve separate treatment in addition to long-accepted technological readiness levels to provide a comprehensive metric for technology and innovation management.

We suggest the implementation of the Dual Use Research of Concern (DURC) framework, originally designed for life sciences, to the domain of generative AI, with a specific focus on Large Language Models (LLMs). With its demonstrated advantages and drawbacks in biological research, we believe the DURC criteria can be effectively redefined for LLMs, potentially contributing to improved AI governance. Acknowledging the balance that must be struck when employing the DURC framework, we highlight its crucial political role in enhancing societal awareness of the impact of generative AI. As a final point, we offer a series of specific recommendations for applying the DURC approach to LLM research.

Various approaches have emerged over the last several decades to meet the challenges and complexities of anticipating and responding to the potential impacts of emerging technologies. Although many of the existing approaches share similarities, they each have shortfalls. This paper takes as the object of its study Anticipatory Ethics for Emerging Technologies (ATE) to technology assessment, given that it was formatted to address many of the privations characterising parallel approaches. The ATE approach, also in practice, presents certain areas for retooling, such as how it characterises levels and objects of analysis. This paper results from the work done with the TechEthos Horizon 2020 project in evaluating the ethical, legal, and social impacts of climate engineering, digital extended reality, and neurotechnologies. To meet the challenges these technology families present, this paper aims to enhance the ATE framework to encompass the variety of human processes and material forms, functions, and applications that comprise the socio-technical systems in which these technologies are embedded.

La technologie moderne crée la capacité de voir et d’agir à l’échelle où, pourvu qu’on ne les aide pas, les sens humains n’ont aucun pouvoir. La combinaison d’une distance infranchissable entre l’homme et la technique avec un réel pouvoir qu’a la technique sur la condition de l’homme, caractérise, non seulement les technologies émergentes comme les nanotechnologies, mais aussi la théologie, qui explore une réalité divine postulée. Nous utilisons cette analogie anthropologique afin d’analyser le fonctionnement des images et noms technologiques dans la société.

Are digital subjects in virtual reality morally equivalent to human subjects? We divide this problem into two questions bearing, respectively, on cognitive and emotional equivalence. Typically, cognitive equivalence does not hold due to the lack of substantialist indistinguishability, but emotional equivalence applies: digital subjects endowed with face or language elicit emotional responses on a par with real-world pleasure, desire, horror, or fear. This is sufficient for projecting moral traits on avatars in the metaverse or on dialog systems based on large language models. Our main case study is a chatbot trained on the chat history between a Canadian man and his deceased fiancée. To demonstrate emotional equivalence and the mechanism of moral transfer, we compare digital devices with the functioning of oracles in a story by Plutarch and in a narrative that draws on the book of Genesis. Finally, we note that, along with the projections of ethical issues, humans also tend to bring real-world solutions of moral conundrums into extended reality. We argue that the lack of cognitive equivalence makes such projections problematic as they lead to overpolicing and a sanitized metaverse.

An operational approach to quantum theory consists in a derivation of its formalism from fundamental postulates with a clear physical meaning. This approach anchors fundamental issues, e.g. causality, continuity or compositionality, in the mathematical study of “post-quantum” models and their comparison with quantum theory. Such non-classical models enable a new understanding of quantum theory and strengthen its philosophical interpretation in the language of information. Quantum theory emerges as a “resource theory”: the non-classical resources, e.g. entanglement or contextuality, can be quantified against the background of alternative models and consequently used to build quantum technological devices. On the philosophical side, it is an open issue whether the information-theoretic approach amounts to a particular informational form of structural realism or delves into the issues of subjectivity and phenomenology of the quantum mechanical observer.

Watermarks should be introduced in the natural language outputs of AI systems in order to maintain the distinction between human and machine-generated text. The ethical imperative to not blur this distinction arises from the asemantic nature of large language models and from human projections of emotional and cognitive states on machines, possibly leading to manipulation, spreading falsehoods or emotional distress. Enforcing this distinction requires unintrusive, yet easily accessible marks of the machine origin. We propose to implement a code based on equidistant letter sequences. While no such code exists in human-written texts, its appearance in machine-generated ones would prove helpful for ethical reasons.

Due primarily to technological advances over the last decade, quantum research has become a key priority area for science and technology policy all over the world. With this manifesto, we wish to prevent quantum technology from running into fiascos of implementation at the interface of science and society. To this end, we identify key stumbling blocks and propose recommendations.

Les agents conversationnels (aussi appeles « chatbots ») sont des systemes numeriques capables d’interagir avec leurs utilisateurs en langage naturel, a l’ecrit comme a l’oral. Aujourd’hui, on en trouve dans de nombreux objets de notre quotidien (smartphones, enceintes connectees, voitures, sites internet) : ils apportent des conseils de sante, facilitent le recrutement, gerent des services apres-vente, forment a distance, aident les personnes vulnerables, etc. L'utilisation de ces systemes ne cesse de croitre. Leur performance grandit aussi tres vite : l’utilisation des « transformeurs » — gigantesques reseaux de neurones tres performants — revolutionne le traitement automatique du langage et tout particulierement la conception des agents conversationnels. Les concepteurs de ces systemes envisagent de nouveaux developpements capables de bouleverser certains aspects de notre vie : des « amis virtuels », des « anges gardiens » ou encore des « deadbots » (chatbots entraines a partir des donnees d’une personne defunte). En France comme au niveau europeen, les agents conversationnels, parce qu’ils s’appuient sur des donnees personnelles, commencent a attirer l’attention des autorites, des regulateurs et du public . Au-dela de ces points reglementaires, se posent de nombreuses questions ethiques. Quels comportements et croyances avons-nous par rapport aux agents conversationnels ? Un chatbot peut-il mentir a son utilisateur ? Les erreurs des agents conversationnels seront-elles plus acceptables que celles d’un etre humain ? Un chatbot doit-il imiter systematiquement l’etre humain ? Pour enrichir sa reflexion et sensibiliser aux questions ethiques, le CNPEN a mis en place un appel a contribution . Il a recu une centaine de reponses individuelles ou collectives. Parmi les differents enjeux d’ethique qu’identifie le comite, les principales tensions concernent le brouillage des frontieres entre une machine et un etre humain, l’imitation du langage et des emotions par les chatbots, ainsi que la capacite des agents conversationnels a manipuler les interlocuteurs humains. L’avis s’adresse aux industriels en proposant de mettre en œuvre dix principes de conception des chatbots. Treize preconisations sont formulees a l’adresse des autorites publiques qui ont la responsabilite d’accroitre l’effort de formation, d’education, mais aussi d’experimentations a l’echelle de la societe. Enfin, l’avis enonce onze questions qui pourraient faire l’objet de nouveaux travaux de recherche. La reflexion sur les enjeux d’ethique des agents conversationnels, relative tant a leurs usages qu’a leur conception, devra etre poursuivie en tenant compte des evolutions culturelles et technologiques.