mediadev library

Sort by:

Direction:

Companion AI: Risiken von sykophantischen und suchtfördernden Designs von KI-Systemen, ihre rechtliche Erfassung und Handlungsempfehlungen
Deep Insights

Ramak Molavi-Vasse’i

Berlin: Zentrum für Digitalrechte und Demokratie (2026), 82 pp.

"Diese Studie untersucht Risiken von KI-Systemen, die im fortlaufenden Dialog mit Nutzern Emotionen erkennen, persönliche Bedürfnisse adressieren und darauf reagieren. Chatbots mit dieser Funktionalität werden als Companion-AI bezeichnet. Sie sind hochgradig personalisiert und treten Nutzern als ... soziales Gegenüber entgegen, das freundschaftliche, romantische oder sexuelle Nähe simuliert. Menschen können zu solchen Systemen eine emotionale Bindung aufbauen. Der Fokus liegt nicht nur auf spezifischen Companion-Anwendungen wie Replika oder Character.AI, sondern auch auf Universalmodellen wie ChatGPT, Claude, Gemini, Grok oder Meta-AI. Diese Systeme werden zunehmend für persönliche, emotionale und beratende Gespräche genutzt. Die Untersuchung kommt zu dem Ergebnis, dass Companion-AI mehrere eng miteinander verbundene Risikodimensionen erzeugt. Psychische und physische Gesundheit: Companion-AI trifft auf eine Gesellschaft, in der Einsamkeit und psychische Belastungen zunehmen, während professionelle Versorgungsangebote knapp sind. Die möglichen Risiken sind klinisch belegt. Die Nutzung kann psychische Belastungen verursachen oder verstärken und in Einzelfällen gravierende Gesundheitsfolgen auslösen. Dokumentiert sind die Verschlimmerung psychotischer Zustände, die Verstärkung depressiver Muster und Angststörungen, suchtartige Bindungen mit Entzugssymptomen sowie die Erosion sozialer Kompetenzen, etwa eine messbar reduzierte Konfliktfähigkeit nach längerer Interaktion mit Companion-AI. Öffentlich bekannt gewordene Vorfälle sind in der Vorfall-Datenbank der Studie dokumentiert. Eingriff in die Privatsphäre: Companion-AI animieren Nutzer kontinuierlich zur Selbstoffenbarung und greifen damit in sensible Gedanken und die intime Gefühlssphäre der Nutzer ein. Zugleich ermöglicht die fortlaufende Interaktion eine zunehmend verdichtete Profilbildung. Entscheidungsautonomie und demokratische Meinungsbildung: Companion-AI können sowohl die Qualität von Informationen als auch die Entscheidungsautonomie der Bürger beeinträchtigen. Die Mechanismen, die Nähe und Vertrauen erzeugen, beeinflussen zugleich die Generierung, Aufnahme und Gewichtung von Informationen. Sykophanz, also die unkritische Bestätigung von Nutzeransichten, beeinträchtigt Genauigkeit und Verlässlichkeit der Antworten messbar. Sprachmodelle werden zunehmend als primäre Quelle der Informationssuche genutzt. Wenn dieselben Systeme Informationen erzeugen, beschaffen und präsentieren, fallen Selektion und Aufbereitung in einer Hand zusammen. Werbe- und interessengeleitete Einflussnahme greift dann nicht mehr nur in einzelne Kaufentscheidungen ein, sondern in die Voraussetzungen öffentlicher Meinungs- und demokratischer Willensbildung." (Executive summary, Seite 5) more

External Link (82)

Relational Futures: Indigenous Sovereignty and the Governance of Artificial Intelligence (AI)

Bronwyn Carlson, Tamika Worrell

Macquarie University (2026), 30 pp.

"Artificial intelligence must be governed as a relational and institutional force, not simply a technical system. Effective policy requires Indigenous leadership, enforceable standards, and a commitment to sovereignty, accountability, and community wellbeing. Failure to act risks embedding inequalit ... more

External Link (100)

Promoting and advancing human rights in global AI ecosystems: The need for a comprehensive framework under international law

Alexander Kriebitz, Caitlin C. Corrigan (eds.)

Globethics Foundation; Responsible AI Network – Africa (RAIN-Africa) (2026), 33 pp.

"In response to far-reaching societal impacts of AI, policymakers are increasingly adopting a human rights lens for AI governance. However, the regulatory landscape remains fragmented and characterized by Western led initiatives as well as inconsistent enforcement. To address these limitations, a un ... ified and binding international framework is urgently needed. The proposed Convention on AI, Data, and Human Rights (drafted as the “Munich Convention on AI, Data, and Human Rights”) is an initiative toward such a framework. Developed with contributions from over 50 global experts, it proposes an international convention to align AI governance with human rights principles. This whitepaper explores the underlying rationale behind the convention, emphasizing its alignment with the norms and principles articulated in the Universal Declaration of Human Rights and the UN Charter. It examines the relevance of safeguards against AI’s misuse in facilitating human rights violations, with particular focus on protecting underrepresented or marginalized groups, including indigenous peoples, and addressing collective rights. It furthermore points to the necessity to codify rights that empower individuals to opt out, be forgotten, seek explanations, and access remedies to maintain adequate human rights standards in the context of AI. The urgency to act arises from the escalating impact of AI and the rapid pace of political and technological advancements. The UN Human Rights Council (UNHRC) is particularly well poised to play a leading role in initiating global discussions on a binding convention for AI governance, grounded in human rights principles. Drawing on its established mandate to uphold human rights globally, the UNHRC has a record of successfully embedding human rights principles into international frameworks, as demonstrated by its endorsement of the UN Guiding Principles on Business and Human Rights and the recognition of the right to a clean, healthy and sustainable environment. Through dedicated action around promoting a convention on AI and human rights, the UNHRC can foster international consensus and proactively shape the future of AI in a manner that prioritizes the values and principles enshrined in the key documents of international human rights law." (Executive summary) more

Download (26)

Standardization of AI Ethics: Stakeholders, Values and Profit

Nicholas Kluge Corrêa, Julia Maria Mönig (eds.)

Bielefeld: transcript Verlag (2026), 161 pp.

"[...] what a certification could look like is only one of the questions that should be asked when discussing ethical certification of AI. Furthermore, how all potentially concerned parties (stakeholders) can be included, which values might be affected, and what this means for those actors who (righ ... tfully) need and want to profit from AI applications were central questions that the philosophy team of the KI. NRW-flagship project „Zertifizierte KI“ (Certified AI) investigated. Further key questions raised in the contributions and throughout the project period include: Can ethics be standardized? Who audits the auditors? How can auditing institutions be aligned with actual societal concerns? What can an assessment look like for a specific area (e.g., human resources)? How are law and ethics intertwined? Which contribution can philosophy provide, now that there is a European regulation on Artificial Intelligence (e.g., in the discussion whether robots in elderly care should be considered as deceptive technology under the AI Act)? Finally, what role do standardization bodies play in ensuring trustworthy AI? Answers to these questions are presented in this book, as well as in the project’s central outcomes: the 'Catalogue of General Ethical Requirements for AI Certification' and the 'Catalogue of Field-Specific Requirements for AI Certification'. The 'Catalogue of General Ethical Requirements for AI Certification' provides a structured framework for trustworthy AI development, identifying overarching ethical requirements for AI certification and outlining six ethical values: fairness, privacy, robustness, sustainability, transparency, and truthfulness. For each value, the project team researched existing tools, metrics, and approaches to help developers, programmers, and other practitioners build trustworthy AI systems. The 'Catalogue of Field-Specific Requirements for AI Certification' offers a list of requirements, an ethical stakeholder analysis, and a discussion of potential value tensions for six exemplary fields: arts, biometric profiling, industry and critical infrastructure, law enforcement, medicine and care and social media, a list of requirements, an ethical stakeholder analyses and a discussion of potential value tensions. The volume at hand collects the contributions to the final conference of the sub-project philosophy of the KI.NRW-Flagship-Projekt “Zertifizierte KI”. (Introduction, pages 11-12) more

External Link (112)

Toward AI Governance Alignment in Africa, Middle East, and Türkiye (AMET) Region

Emma Ruttkamp-Bloem, Fola Adeleke, Jake Effoduh, Kebene Wodajo, Nagla Rizk, Olubayo Adekanmbi, Rachel Adams, Rami Alkarmi, Seani Rananga, Seydina Ndiaye, Shikoh Gitau, Tami Koroye

Global Center on AI Governance (GCG) (2026), 54 pp.

"The analysis begins by identifying the core “centric approaches” that guide Artificial Intelligence (AI) national strategies. These include security-anchored models driven by state control, innovation-led frameworks that prioritise competitiveness, rights-based regimes centred on fundamental pr ... otections, and developmental approaches focused on public service delivery and technological catch-up. Each approach expresses incentives, risk perceptions and administrative capacities. Out of these foundational choices, there are different groupings. Various regions within AMET are adopting recognisable patterns of regulation that reflect their political economies, institutional structures and geopolitical positions. These groupings are not static categories, but evolving alignments shaped by domestic pressures, supply-chain dependencies and shifting global power dynamics. Across all regions, a set of cross-cutting issues consistently shapes the regulatory landscape. Data protection and interoperability remain central, alongside growing demands for accountability, explainability and robust auditing. Standards and certification are emerging as tools for both trustbuilding and geopolitical influence. Sector-specific rules in finance, health, education and public administration are increasingly decisive. Capacity constraints, fiscal limits and uneven institutional strength continue to define what is realistically possible for many governments. Geopolitical tensions and cloud concentration reinforce questions around sovereignty, infrastructure control and long-term strategic autonomy. This report translates these insights into practical implications for states, regional bodies and technology firms. It highlights workable pathways for reconciling innovation with safety, managing cross-border data flows, and building resilient digital infrastructure. It also identifies areas where architecture choices such as regional cloud centres, culturally calibrated testing layers and tiered sovereignty models can support both regulatory goals and developmental priorities. The recommendations emphasise flexible alignment, cooperation through standards, and the need for shared mechanisms that accommodate political diversity while reducing fragmentation. Taken together, the framework provides a stable and operational way of understanding global AI regulation. It clarifies the interests driving different jurisdictions, the pressures that shape convergence or divergence, and the strategic considerations that technology firms and policymakers must navigate as the regulatory environment continues to evolve." (Executive summary) more

Download (8)

Inteligencia artificial y desarrollo humano integral

Grupo de Trabajo de Frontera Tecnológica del Celam

Bogotá: Consejo Episcopal Latinoamericano y Caribeño (CELAM) (2026), 86 pp.

" ¿Cómo puede la IA contribuir al desarrollo humano integral y solidario? ¿Cómo podemos hacer uso de los desarrollos tecnológicos de tal modo que contribuyan al bien común, considerando todas las dimensiones de la persona (“desarrollo integral”) y sin excluir a nadie (“desarrollo solidar ... more

Download (42)

Digital Media Shadowing Democracy: Technology, Communication, and Power

Auksė Balčytienė, Péter Bajomi-Lázár, Helena Sousa (eds.)

Gothenburg: Nordicom, University of Gothenburg (2026), 282 pp.

"The book’s message is brief: Liberal democracies are at risk. Democratic decay has numerous causes, but technological innovations and profit-driven dynamics are shaping societal relations in ways that intensify disagreements and polarisation, undermining informed citizenship. The growing number o ... more

External Link (103)

AI in Africa: A Landscape Study

Rachel Adams, Fola Adeleke, Leah Junck, Samule Segun, Ayantola Alayande, Selamawit Abdella, Mark Gaffley, Fabio Andres Diaz Pabon, Funbi Salami

Global Center on AI Governance; African Observatory on Responsible AI (2025), 44 pp.

"Though comprehensive AI regulation is still nascent in Africa, there is an increased focus on national AI strategies, ethical guidelines and international collaborations. Over 40 countries have already established data protection and privacy laws, creating opportunities for harmonised regulations a ... nd a supportive environment for AI innovation. Key continental governance instruments and initiatives also include the African Union’s AI Strategy, the African Continental Free Trade Area and the African Commission on Human and Peoples’ Rights. This report found that the increasing application of AI in key sectors across Africa, such as health, education, finance, agriculture, public services and the environment, necessitates a collaborative approach to ethical AI development and use. This requires addressing persistent infrastructural gaps, developing regional regulatory frameworks, investments in AI research and capacity building efforts as well as promoting multi-stakeholder AI governance initiatives. Key strategies identified include joint investment and funding, pooling resources and supporting sustainable and fair open-source projects, which will help direct efforts toward innovation and regional independence. Open Source AI models are particularly important in Africa where local tech startups can leverage global open-source initiatives to build context-specific solutions. However, concerns regarding data ownership, privacy and ethical AI deployment call for more robust policies for open source AI on the continent. To foster responsible AI innovation, African countries should establish comprehensive AI governance frameworks while adhering to existing laws as well as strengthen regulatory institutions to provide oversight over AI technologies. Regulatory sandboxing also offers a viable approach to enable sage experimentation with appropriate oversight. Regional and international cooperation as well as leveraging public-private partnerships can promote funding for expanding AI research and development initiatives, resolve talent shortages and promote infrastructure development. Africa has a unique opportunity to leverage AI, along with its resources and local young talent, for socioeconomic development. However, to ensure inclusivity and sustainable innovation, it is essential to address infrastructure challenges and establish regulatory frameworks that protect societal wellbeing. Embracing collaborative efforts, including open source initiatives, regional and international partnerships are critical for achieving the continent’s ambitious goals without leaving anyone behind." (Executive Summaty, page 3) more

Download (63)

An Autonomy-Based Classification: AI Agents, Liability and Lessons from the Automated Vehicles Act

Lisa Soder, Julia Smakman, Connor Dunlop, Oliver Sussman

Berlin: Interface; Ada Lovelace Institute (2025), 28 pp.

Download (55)

The AI Policy Playbook: Navigating AI Policy Development through an African and Asian Lens

Aubra Anthony, Deshni Govender, Nadine Dammaschk, Golo Rademacher, Kim Willet, Gustavo Fonseca Ribeiro, Sheila Kibughi

Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) FAIR Forward; AI Policymaker Network (2025), 70 pp.

"The AI Policy Playbook offers a practical guide for governments in Africa and Asia to develop responsible and context-specific AI policies. Drawing from real-world policymaking experiences in seven countries, Ghana, Kenya, Rwanda, South Africa, Uganda, India, and Indonesia, the playbook shares less ... more

Download (26)

Inclusive and Secure Artificial Intelligence: A Global Perspective on Policy and Technical Developments
Deep Insights

Saiph Savage, Lili Savage

Stuttgart: Institut für Auslandsbeziehungen (ifa) (2025), 84 pp.

"As AI systems are developed and deployed, they often reflect and reinforce dominant cultural norms, marginalizing non-Western languages, epistemologies, and communities. This report explores the systemic risks associated with algorithmic bias, digital colonialism, and cultural homogenization, while ... more

Download (35)

The State of AI in Africa Report

Nairobi: Strathmore University, Centre for Intellectual Property and Information Technology Law (CIPIT) (2025), 67 pp.

"The State of AI in Africa Report 2025 offers a critical reflection on how AI ecosystems across the continent have evolved since 2023, highlighting advances in infrastructure, regulatory frameworks, capacity building, and sectoral innovation, while also surfacing enduring structural disparities. Thi ... s report reveals a continent that is no longer on the margins of AI development but actively reshaping how AI is conceived, deployed, and governed. Kenya, South Africa, Nigeria and Ghana have made notable strides in anchoring AI within national digital strategies, strengthening local innovation ecosystems, and investing in foundational infrastructure such as data centres and broadband connectivity. The growing presence of indigenous datasets, grassroots NLP initiatives, and AI models tailored to African realities reflects a broader shift from consumption to creation. Yet this progress remains uneven. While pockets of excellence are emerging, many regions continue to grapple with limited access to reliable electricity, fragmented regulation, data governance challenges, and critical skill shortages. Central Africa, in particular, still faces infrastructural and institutional constraints that hinder meaningful participation in the continental AI space. This report also documents how AI is being embedded in key sectors; agriculture, health, law, and the creative economy to address uniquely African challenges – from crop failure and disease surveillance to language preservation and cultural expression. These use cases affirm the value of locally informed, contextsensitive AI innovation. On the governance front, the adoption of the AU-AI Continental Strategy and the development of national AI policies signal a growing awareness of the need for harmonised, responsible and inclusive AI development. However, effective implementation remains a central concern, particularly around enforcement mechanisms, ethical guardrails, and regional coordination. This report presents a nuanced picture of AI in Africa: one of ambitious experimentation, uneven capacity, and deep potential. Africa’s future in AI depends not only on technological advancement but on the continent’s ability to embed equity, local relevance, and sustainability at the heart of its AI agenda." (Executive summary) more

Download (67)

Guía de implementación de protección de datos y uso responsable de inteligencia artifical en Bolivia

Estefani Cabrera Rojas, Nicoel Angel Sanchez Rojas

La Paz: InternetBolivia; Consorcio InDiCo Global (2025), 63 pp.

"En Bolivia, aunque no existe una ley integral de protección de datos personales, el marco constitucional y normativas sectoriales reconocen el derecho a la privacidad, lo que refuerza la necesidad de avanzar hacia políticas institucionales de cumplimiento voluntario, con base en estándares inter ... more

Download (20)

LAO P.D.R. Artificial Intelligence Readiness Assessment Report

Paris: UNESCO (2025), 72 pp.

"• In the legal context, Lao PDR’s regulatory framework for AI is still nascent. While the government is drafting a National AI Strategy, there are currently no laws addressing core AI issues such as transparency, accountability, or algorithmic redress. Existing frameworks like the Cybercrime La ... w, Data Protection Law and the Digital Government Decree cover general data and computing use but lack AI-specific provisions. AI is referenced in recent constitutional amendments and draft digital transformation policies, but there’s room for improvement in transforming those acknowledgements to clear legal mechanisms to protect citizens and bolster innovation. • For the sociocultural domain, despite progress in gender inclusivity, rural communities in Laos still face limited access to digital technologies due to infrastructure gaps, affordability, and low digital literacy. Successful collaborations with international organizations highlight what’s possible but also stress the need for ongoing investment in local capacity, outreach, and trust-building. Digital initiatives in health and efforts addressing concerns about technology’s environmental impact are evident, but they have not been specifically focused on AI. Digital transformation must also prioritize cultural preservation, including the development of culturally-aware datasets and NLP tools. Currently, most Lao language datasets and NLP tools are lagging participation from native Lao researchers, risking linguistic sovereignty and underrepresenting local dialects, norms, and values. • AI talent and innovation development in Laos needs strong support from its education and research ecosystems. University departments like Computer Science and Software Engineering are small and often siloed, with little crossdisciplinary collaboration. Research output needs some improvement. Promising developments in AI/ICT curricula in K-12 level are welcome, although their effectiveness must be closely monitored. • Laos’s digital economy remains small, constrained by a large informal sector, limited digital infrastructure, and low ICT investment, factors that hinder AI adoption and innovation. While some private companies in sectors such as telecom, banking, logistics, and manufacturing are beginning to use AI tools, and urban populations are gradually adopting app-based lifestyles, broader progress is limited by lagging human capital pipelines, minimal domestic investment, and an underdeveloped startup ecosystem. Foreign-developed platforms still continue to dominate the market, which if continued can limit local developers from developing capabilities and use cases. Aside from that, most Foreign Development Investment (FDI) remains concentrated in resource extraction, and budget shortfalls undermine the sustainability of e-government initiatives. • There is a critical need for a strong commitment to funding both the hard and soft infrastructure that supports AI. This includes enhancing cybersecurity capacity, reliable networks, and data centers and ensuring affordable access to computing infrastructure for AI practitioners and stakeholders." (Executive summary) more

Download (44)

Just Hiring! How to counter discrimination in AI-assisted recruiting at the governance level. A toolkit for policymakers

M. Daviet, A. Mueller, N. Baranowska, G. Meijers, C. Castillo, J. Saldivar, A. Gatzioura

FINDHR Fairness and Intersectional Non-Discrimination in Human Recommendation (2025), 67 pp.

"In this Toolkit, we offer introductory information and concrete recommendations on how to mitigate the risk of discrimination in algorithmic hiring systems on a governance level. This Toolkit addresses policymakers and people working in regulatory bodies, as well as civil society organisations inte ... more

Google Scholar (142)

WorldCat (144)

Intelligence artificielle et robotique autonome: Enjeux et défis contemporains dans l’espace francophone

Etienne Damome, Moustapha Ben Ismïila Diaby (eds.)

Paris: L'Harmattan (2025), 358 pp.

"Cet ouvrage propose une analyse nuancée des transformations induites par l'IA et la robotique autonome dans l'espace francophone. Il insiste sur son potentiel à façonner durablement les sociétés contemporaines. S'appuyant sur des exemples concrets et des études de cas, l'ouvrage fournit un ca ... more

Download (59)

Companion Document: The Guidelines for the Governance of Digital Platforms and Generative Artificial Intelligence

Paris: UNESCO (2025), 34 pp.

"The Guidelines for the Governance of Digital Platforms outline a set of responsibilities, duties, and recommendations relevant to the governance of generative AI. These include fostering an enabling environment for freedom of expression and access to information; establishing independent governance ... more

Download (58)

Governing the Digital Society: Platforms, Artificial Intelligence, and Public Values

José van Dijck, Karin Van Es, Anne Helmond, Fernando van der Vlist (eds.)

Amsterdam: Amsterdam University Press (2025), 261 pp.

"Governing the Digital Societypresents a range of governance approaches, focusing on online platforms, artificial intelligence, and the public values that underpin these technologies. The authors position themselves at the forefront of their disciplines, offering perspectives from law, critical data ... more

Download (27)

Pathways to Inclusion: Advancing Civil Society’s Role in AI Governance: Evidence from and for the Global Majority

Paola Gálvez Callirgos

Geneva: Globethics Publications (2025), 58 pp.

"This policy brief investigates the role of civil society on AI Governance with particular attention to the Global Majority. The research question explores whether civil society serves as an effective counterbalance to concentrated power in the AI ecosystem to ensure AI systems reflect diverse publi ... more

Download (39)

Harnessing the Power of AI: Safe & Ethical Integration. Report of the AI Safety Dialogue at Global Ethics Forum 2025

Geneva: Globethics (2025), 27 pp.

"AI presents both significant opportunities and serious risks. For developing countries, AI offers immense potential for advancing learning, research, and development. At the same time, it poses complex challenges that demand collective action. Diplomats and policymakers recognize that global cooper ... ation and multilateralism are essential. No single nation can manage these challenges alone. International and policy initiatives, including the Global Digital Compact and the AI Safety Summit series, have established broad frameworks for addressing these issues, providing global ideas that now require practical implementation. This report emerges from that momentum. Through Track II Diplomacy for AI Global Safety dialogue, Globethics has built a community of practice bringing together cross-disciplinary experts from across geopolitical contexts to explore AI-related risks and identify existing harms; particularly those affecting the Global Majority. Two principal categories of risks have been identified: 1. Risks triggered by advanced AI systems versus those resulting from the intended or unintended use of AI by human actors. 2. Future risks versus those already manifesting today. The words of UN Secretary-General António Guterres resonated across the sessions: “the question is not whether AI will influence international peace and security, but how we will shape that influence.” While discussions often swing between hype and horror, this report intends to shape pathways towards hope and trust. Trust, essential for AI adoption, is vital for both the private sector and civil society. Likewise, it emphasizes the importance of human rights, without which there can be no sustainable development or peace. At the core of these global AI Safety discussions was the need to advance concrete mitigation strategies, including: capacity building; addressing existing harms; developing operational frameworks; ensuring intersectional, interdisciplinary, cross-sectoral, multistakeholder participation. A coordinated approach to global safety, diplomacy, and multilateral governance is imperative. Experts agreed that tangible measures, rather than fragmented initiatives and principles, are needed. By coming together across regions, disciplines, and cultures, experts reaffirmed their commitment to stewarding technology for the common good. The dialogues converged into three critical priorities to transform today’s innovation into meaningful societal progress: Strengthening inclusive, representative participation using empowerment and literacy measures for people and communities to shape AI futures; building systems of trust, verification and ethical responsibility across the AI lifecycle marked by due diligence, effective reporting and timely safeguards by design; translating values into action, ensuring coordination and accountability from design to deployment." (Executive summary, page 3) more

Companion AI: Risiken von sykophantischen und suchtfördernden Designs von KI-Systemen, ihre rechtliche Erfassung und Handlungsempfehlungen Deep Insights

Inclusive and Secure Artificial Intelligence: A Global Perspective on Policy and Technical Developments Deep Insights

Companion AI: Risiken von sykophantischen und suchtfördernden Designs von KI-Systemen, ihre rechtliche Erfassung und Handlungsempfehlungen
Deep Insights

Inclusive and Secure Artificial Intelligence: A Global Perspective on Policy and Technical Developments
Deep Insights