Social robots are increasingly present in homes, schools, and care settings, but in regulations and international standards they are typically classified as mobile service robots (MSRs). What qualifies as an MSR varies across jurisdictions and standards, creating definitional ambiguity with practical safety consequences: for example, Paro, the robotic seal, is classified as a medical device in the U.S. but treated as non-medical in the EU, complicating which safety requirements apply. This problem is amplified by ongoing changes in standards, including ISO 13482’s shift from ‘personal care robots’ (2014) to broader ‘service robots’ in the 2024 draft. To support clearer hazard analysis, testing, and oversight, we systematically review MSR definitions in research and standards and map resulting misalignments. Using PRISMA, we synthesize recent literature and compare it with ISO 13482:2014, the revised ISO/DIS 13482:2024, and ISO 8373:2021 on robotics vocabulary. Our results show that vulnerable groups, especially children, older adults, and pregnant women, remain insufficiently protected; emotional and cognitive risks are rarely specified; and standards continue to emphasize physical hazards. Many robots described in the literature do not clearly align with ISO MSR criteria, creating uncertainty for hazard analysis and conformity assessment. Based on these findings, we propose a refined MSR definition incorporating diverse embodiments, services, and social interaction.

Social media platforms rely on Gender Classification Systems (GCSs) to infer users’ gender from behavioral and demographic data, often without explicit consent. These systems optimize targeted advertising and user engagement but introduce significant ethical and regulatory concerns related to algorithmic bias, privacy, data governance, and accountability. Our study presents a large-scale survey (N=1642) analyzing the accuracy and implications of X’s (formerly Twitter) gender inference mechanisms that reveal systemic biases that disproportionately impact marginalized communities. Our findings indicate that men are less likely to experience misclassification than women. Furthermore, LGBTIQ+ individuals and those with non-conforming gender expressions face significantly higher risks of algorithmic misidentification. These results expose critical vulnerabilities in automated profiling systems and highlight the limitations of reductionist, binary technical frameworks applied to the inherently complex and fluid nature of gender identity. Our work underscores the urgent need for improved information management practices involving GCSs, emphasizing compliance, transparency, and user agency. By addressing these challenges, platforms can better align with evolving regulatory frameworks and societal expectations regarding data responsibility, fairness, and inclusion. These insights contribute to the growing imperative for inclusive, rights-based algorithmic governance across social media platforms.

The rapid advancement of service robotics has outpaced regulatory frameworks, leading to gaps and inconsistencies that hinder effective governance. While evidence-based policymaking is well-established in health and consumer protection fields, robotics regulation remains fragmented and reactive. This paper proposes Science for Robot Policy, a structured, evidence-driven model that bridges the disconnect between robotics innovation and regulatory adaptation. Using a Constructive Research Approach, the model integrates scientific experimentation, stakeholder engagement, and knowledge brokering to generate policy-relevant data and transform it into actionable regulatory insights. The model follows a five-step process, beginning with risk identification and prioritization, followed by controlled experimentation in simulators, testing zones, living labs, and real-world markets. The ambition is that insights generated are then translated into policy-relevant information and further refined into knowledge for policymakers, ensuring that empirical evidence informs that robotics regulation is dynamic, anticipatory, and informed. This approach contributes to ongoing discussions on science-for-policy methodologies and fosters iterative regulatory refinement in service robotics. If successful, such a model could allow policymakers to address emerging risks proactively, reduce regulatory uncertainty, enhance user safety, and promote responsible robotics innovation by embedding scientific insights into the policy cycle.

This article analyses ideas to use AI-supported systems to counter ‘cognitive warfare’ and critically examines the implications of such systems for fundamental rights and values. After explicating the notion of ‘cognitive warfare’ as used in contemporary public security discourse, the article describes the emergence of generative AI tools that are expected to exacerbate the problem of adversarial activities against the online information ecosystems of democratic societies. In response, researchers and policymakers have proposed to utilize AI to devise countermeasures, ranging from AI-based early warning systems to state-run content moderation tools. These interventions, however, interfere, to different degrees, with fundamental rights and values such as privacy, communication rights, and self-determination. This article argues that such proposals insufficiently account for the complexity of contemporary online information ecosystems, particularly the inherent difficulty in establishing causality and attribution. Reliance on the precautionary principle might offer a justificatory frame for AI-enabled measures to counter ‘cognitive warfare’ in the absence of conclusive empirical evidence of harm. However, any such state intervention must be based in law and adhere to strict proportionality.

Complex technologies such as Artificial Intelligence (AI) can cause harm, raising the question of who is liable for the harm caused. Research has identified multiple liability gaps (i.e., unsatisfactory outcomes when applying existing liability rules) in legal frameworks. In this paper, the concepts of shared responsibilities and fiduciary duties are explored as avenues to address liability gaps. The development, deployment and use of complex technologies are not clearly distinguishable stages, as often suggested, but are processes of cooperation and co-creation. At the intersections of these stages, shared responsibilities and fiduciary duties of multiple actors can be observed. Although none of the actors have complete control or a complete overview, many actors have some control or influence, and, therefore, responsibilities based on fault, prevention or benefit. Shared responsibilities and fiduciary duties can turn liability gaps into liability overlaps. These concepts could be implemented in tort and contract law by amending existing law (e.g., by assuming that all stakeholders are liable unless they can prove they did not owe a duty of care) and by creating more room for partial liability reflecting partial responsibilities (e.g., a responsibility to signal or identify an issue without a corresponding responsibility to solve that issue). This approach better aligns legal liabilities with responsibilities, increases legal certainty, and increases cooperation and understanding between actors, improving the quality and safety of technologies. However, it may not solve all liability gaps, may have chilling effects on innovation, and may require further detailing through case law.

Gender classification systems (GCSs) on social media platforms, such as X (formerly Twitter), infer users’ gender for targeted advertising and personalization. However, these systems rely on binary classifications that fail to capture gender diversity, often leading to misclassification (i.e. misgendering). This study is a comprehensive analysis of algorithmic misgendering and its impact on user perceptions through a large-scale online global survey (N = 1523). In the first stage, we assess the prevalence of misgendering on X by analyzing the accuracy of inferred gender classifications. Our findings reveal that women and LGBTQ+ users are disproportionately misclassified, which highlights structural biases in gender inference systems. In the second stage, given that the emotional and social consequences of algorithmic gender inference remain underexplored, we examine how users perceive and respond to misgendering. Using ordinal logistic regression models, we find that individuals who experience misgendering report significantly more aversion to X's gender policies. Furthermore, increased platform engagement is linked to stronger opinions on gender inference, reducing neutrality toward these systems. Beyond these findings, our study also reveals how opaque gender classification is, as many users struggle to locate, understand, or challenge their inferred gender within X's interface. This lack of transparency raises concerns about agency, algorithmic literacy, data protection, and fairness. Based on our work, we suggest regulatory measures to ensure greater transparency and user control over gender classification, contributing to ongoing debates on algorithmic discrimination and inclusive AI governance on current and future social media platforms.

Over the past two decades, socially assistive (SARs) or interactive (SIRs) robots have been developed rapidly due to their beneficial uses in elderly care, rehabilitation, and education. Given their multiple embodiments and contexts of use, however, defining what these robots are remains a difficult task, which further challenges understanding which legal safeguards developers need to follow to ensure a safe human-robot interaction (HRI). Establishing legislation that adequately frames the issues is complex if these concepts remain confusing. Despite pioneer efforts to characterize what these robots are in international standards and the literature, there is currently no consensus on which legal category they are, and, therefore, related problems are covered unevenly in different pieces of legislation. Following a systematic review, we analyzed 1,359 works (out of 3,446) to clarify definitions, categories, and functionalities of SARs and SIRs to establish a baseline for understanding and regulating these robots. The first results show that the ISO 13482:2014 definition of Mobile Service Robots (MSRs), the formal name for SARs, is incompatible with the current literature. Moreover, more consensus on what qualifies as assistive or interactive under this technology is needed to determine the regulation and safeguards to mitigate the issues these robots entail for users

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

Drones are becoming a more familiar sight in the skies. This growth has potential implications for all aspects of society. To ensure safe, secure and sustainable drone activities, regulatory authorities have been revoking, amending or introducing laws. The primary focus has been on aviation safety rules for the design, manufacture and operation of these unmanned aircraft. These rules do not directly address other domains. This paper assesses how the EU’s drone-specific safety-based rules interact with the EU’s AI legal framework that is currently being developed. While drone rules do not address AI directly, they do indirectly as, for example, autonomous drone operations are permitted in the single European sky. The general AI legal framework also applies to drone activities, despite such not being mentioned. This paper brings together the sector-specific EU drone rules and the EU’s generalAI legal framework to understand what the term ‘AI’ means within the context of drone operations, and assesses whether there are any inconsistencies, contradictions, overlaps or lacunae. This will determine the applicability and suitability of applying the general AI legal framework to a specific technology (drones) and sector (aviation), and whether the drone rules can contribute to the regulation of AI more generally.

According to the fairness principle in Article 5.1a of the EU General Data Protection Regulation (GDPR), data controllers must process personal data fairly. However, the GDPR fails to explain what is fairness and how it should be achieved. In fact, the GDPR focuses mostly on procedural fairness: if personal data are processed in compliance with the GDPR, for instance, by ensuring lawfulness and transparency, such processing is assumed to be fair. Because some forms of data processing can still be unfair, even if all the GDPR's procedural rules are complied with, we argue that substantive fairness is also an essential part of the GDPR's fairness principle and necessary to achieve the GDPR's goal of offering effective protection to data subjects. Substantive fairness is not mentioned in the GDPR and no guidance on substantive fairness is provided. In this paper, we provide elements of substantive fairness derived from EU consumer law, competition law, non-discrimination law, and data protection law that can help interpret the substantive part of the GDPR's fairness principle. Three elements derived from consumer protection law are good faith, no detrimental effects, and autonomy (e.g., no misleading or aggressive practices). We derive the element of abuse of dominant position (and power inequalities) from competition law. From other areas of law, we derive non-discrimination, vulnerabilities, and accuracy as elements relevant to interpreting substantive fairness. Although this may not be a complete list, cumulatively these elements may help interpret Article 5.1a GDPR and help achieve fairness in data protection law.

The EU General Data Protection Regulation (GDPR) contains several data subject rights, but for many of these rights it is not entirely clear how they should work in practice, especially in digital environments. Most data subject rights apply to personal data obtained directly or indirectly from the data subject. This is often personal data that data subjects already are familiar with, i.e., things they already know about themselves. Unclear, however, is to what extent ascribed personal data, such as inferred data and categories or profiles in which data subjects are placed by data controllers, are within the scope of these rights. Such ascribed personal data often concerns novel information, generated by data controllers, and includes insights into how controllers view and assess them, which may have practical and legal impact on data subjects. Given these characteristics, the ascribed personal data may be much more interesting to data subjects, so it appears beneficial, from the policy perspective, to have this novel information included in the scope of data subject rights. If data subject rights do not apply to inferred data and profiles, invoking these rights is unlikely to be informative and provide meaningful information for data subjects, particularly in complex, digital environments. However, if data subject rights do apply to inferred data and profiles, the scope of these rights may be hard to delineate and they may quickly interfere with rights and freedoms of others, including trade secrets of data controllers and privacy rights of other data subjects. In this article, we investigate the implications of applying data subject rights to inferred data and profiles. For each data subject right in the GDPR, we assess which types of personal data could and perhaps should be in scope, based on grammatical and teleological legal analyses as well as practical considerations. While the area of data subject rights received significant academic attention in the past years, our article contributes to the discussion by providing a systematic, holistic framework to consider the scope of the rights in relation to ascribed data.