commbox Artificial Intelligence & Algorithms

Artificial Intelligence & Algorithms

Filter

375

Sort by:

Direction:

Download (94)

Indic Voice Technologies for an Inclusive Digital India: Toolkit for Developers

Nihar Desai, Sujith P., Harleen Kaur, Dona Mathew, Jyotsna Jayaram, Padmavathi Prasad, Thomas Vallianeth

Bhashini; Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ); FAIR Forward – Artificial Intelligence for All (2026), 70 pp.

"This toolkit sets out a layered, lifecycle-oriented approach to building inclusive and robust speech artificial intelligence (AI) systems. It brings together strategies for diverse and representative data collection, linguistically informed model training, rigorous quality control, and deployment o ... more

External Link (187)

Encyclical Letter Magnifica Humanitas of His Holiness Pope Leo XIV on Safeguarding the Human Person in the Time of Artificial Intelligence

Leo XIV (Pope)

Vatican City: Holy See (2026)

"Divided into five chapters, Magnifica humanitas has an underlying premise: technology is not “a force antagonistic to humanity” (paragraph 4), nor is it “inherently evil” (9). However, “technology is never neutral, because it takes on the characteristics of those who devise, finance, regu ... late, and use it. Therefore, Pope Leo XIV appeals for people to build “for the common good” and to “remain human,” following a courageous mentality of shared responsibility and communion, so that the world “will come to recognize the human heart as the place where God desires to dwell” (16). [...] An ethical code for AI: The third chapter—Technology and Dominance. The Grandeur of Humanity in Light of the Promises of AI stresses the need to approach artificial intelligence with vigilance. Pope Leo warns about the “technocratic paradigm” already denounced by Pope Francis and how it can require that every choice be dictated exclusively by measuring efficiency and profits (92). On the contrary, the most powerful technology is not necessarily the best. AI can imitate and simulate the person, but it does not possess a moral conscience, empathy, or affective, relational or spiritual capabilities. The Pope urges clarity about responsibilities and accountability at every stage of the development process, focusing on adequate AI policies and legal frameworks, independent oversight, and user education. Above all, Pope Leo calls for an ethical code subject to shared standards of social justice, because “a more moral AI is not enough if that morality is determined by a few” (107). Nor, he adds, should the environmental impact of new technologies be overlooked, since they require large quantities of energy and water, affecting Creation (101). Disarming AI: AI must be “disarmed,” Pope Leo XIV continues, in order to free it from the mentality of military, economic, and cognitive competition. “To disarm means discrediting the assumption that technical power automatically confers the right to govern,” he says. “To disarm does not mean rejecting technology, but preventing it from dominating humanity” (110). He devotes ample space to a critique of transhumanism and posthumanism, which interpret progress as the overcoming of human limits. Instead, limitations are not defects to be eliminated, but a constitutive dimension of the human person, because it is in fragility and finitude that relationship and openness to God and to others mature. He says we must remember that “humanity flourishes not despite limitations, but often through them” (118). Technological progress without regression of the heart: Pursuing technological innovation at the expense of eliminating human limitations, he says, would cause an anthropological regression. “Humanity—in all its grandeur and woundedness—must never be replaced or surpassed,” he says. Technology can alleviate humanity’s sufferings and open new possibilities, but it must not deny the essence of humanity, which is our “capacity for relationship and love” (126). In the face of AI, says the Pope, “the true alternative is not between enthusiasm and fear, but between two paths of development: a progress that serves individuals and peoples, or a progress that subjects them to the mentality of power” (129). An ecology of communication and the centrality of schools: In the fourth chapter—Safeguarding Humanity at a Time of Transformation. Truth, Work, Freedom—the Pope calls for an “ecology of communication” based on truth. He urges transparency in how content is selected, protection of personal data, serious journalism founded on argumentation and verification, a new awareness in the “proper and critical” use of digital tools, and the integration of different forms of knowledge. The Church must also embody transparent and honest communication, especially in cases of injustice and abuse. The Pope also appeals for a renewed educational alliance, so that the “desire to ask questions” may not be extinguished in young people by perfect machines that make human thought seem useless (140). Pope Leo XIV, therefore, calls for renewed attention to schools as places where people learn to “seek and love the truth” (147). The dignity of work: In the “fourth industrial revolution” represented by the digital transition, the Pope emphasizes the importance of protecting the dignity of work by designing systems centered on the person and not only on performance. “The ‘new ways’ of working are not necessarily better,” he writes, “while AI promises to boost productivity by taking over mundane tasks, it frequently forces workers to adapt to the speed and demands of machines, rather than machines being designed to support those who work” (150). Technology can certainly free human beings from burdensome or repetitive tasks, but it must not lead to unemployment in the name of reducing costs and increasing profit. In this regard, the Pope expresses his hopes for a renewal of labor organizations (155). Peace and development: Pope Leo then notes the need to move beyond GDP as the measure of a country’s level of development, focusing instead on the dignity of work, shared prosperity, the reduction of inequalities, and environmental protection. Finance, he says, must focus on the development, creation, and evolution of work (159-160). Following in the footsteps of Pope St. Paul VI, the encyclical underscores the interdependence between peace and development. It calls for international cooperation capable of defining shared strategies, especially in favor of the most vulnerable countries and groups, because prosperity contributes to peace “only if it is widespread, inclusive, and sustainable” (163). The family, “primary social good”: The Pope upholds the role of the family, saying it is founded on the stable union between a man and a woman. The family is a “primary social good” and the “fundamental and irreplaceable cell of every community organization” (165), which must be supported, including through labor policies that favor stability and humane rhythms, so as to protect society’s ability to “build the future.” ‘Architecture of visibility’ and risks for freedom: Pope Leo XIV then discusses the theme of human freedom in an age when digital platforms are designed to capture users’ time and exploit their vulnerabilities. He reiterates the need to strengthen each person’s interior freedom, while also confronting the risk of social control arising from the mass collection of data and the use of algorithmic systems. Profiling, predicting, and directing behavior, he says, is “a new form of power” (171) that risks discriminating against the weakest. The Pope particularly criticizes the “architecture of visibility,” which amplifies only what is visible and shapes opinions. New forms of slavery and colonialism: AI also generates new forms of slavery, such as that of the “scarred, injured, and worn down” bodies (173) of those who work in the extraction of the “rare earth elements” needed for technology. Therefore, the Pope upholds the importance of combating new forms of slavery as another “decisive test for ethical discernment” in the digital transformation. Pope Leo XIV stresses that “the Church renews her firm condemnation of every form of slavery, trafficking, and commodification of persons” and he underscores that to not react or to tolerate grave violations of human dignity means becoming accomplices to them. At the same time, the Pope “sincerely asks forgiveness” for the delay with which the Church in the past condemned “the scourge of slavery” (174-176). The encyclical also refers to the vital information—for example, on health and demographics—used to guide economic strategies. He calls this a new face of colonialism that turns personal lives into exploitable information, making the digital environment a “space of exploitation” (178-179). Overcoming ‘just war’ theory: In the fifth chapter—The Culture of Power and the Civilization of Love—Pope Leo XIV turns his gaze to war, saying “the digital revolution is changing the nature of conflict.” The Pope calls for an ethical approach, without which decisions about the life and death of persons will become increasingly impersonal due to a use of force regarded as an “immediate and viable option” (182-183). At the root of all this is a “culture of power” that normalizes war and rehabilitates it as an “instrument of international politics,” favoring rearmament. Today, he says, public opinion is burdened by polarizing media narratives, as well as by “a disconcerting loss of historical memory,” which leaves people without a long-term vision (191). Consequently, he says, peace today is no longer understood as a task to be undertaken, but as an interval between conflicts. For this reason, Pope Leo XIV reiterates that—while preserving the right to legitimate defense in the strictest sense—the theory of “just war” must be overcome, and dialogue, diplomacy, and forgiveness must instead be promoted (192). Algorithms do not make war morally acceptable: Pope Leo expresses criticism of the growth in the arms industry, the nuclear arms race, and the emergence of new armed actors—including jihadist groups—who aim to perpetuate conflicts as a source of power and profit. He also warns against the use of weapons linked to AI, because “there is no algorithm that can make war morally acceptable.” The Pope writes, “AI does not remove the intrinsic inhumanity of conflict; indeed, it can only bring about conflict more quickly and render it more impersonal, lowering the threshold for resorting to violence, transforming defense into threat prediction and thus reducing victims to data. In this way, it will accustom us to the idea that violence is inevitable and needs only to be optimized.” The Pope urges strict ethical limits, which are shared at the international level and based on personal responsibility and the protection of civilians. “Any technology that facilitates attacks without seeing the face of human beings lowers the moral threshold of conflict” (199). Crisis of multilateralism: The culture of power also emerges from the crisis of multilateralism and the emergence of “a disorderly and conflict-ridden multipolarism” with a prevailing sense of mistrust (201). The Pope laments that the rule of law has been replaced by the law of the strongest, while the logic of power prevails over peace building and institutions established to safeguard the common destiny of peoples have now been weakened. In this regard, the Pope hopes for “profound reforms” of the UN that can overcome the current crisis of values in favor of the common good (226). An irresponsible Realpolitik: The encyclical notes that today “hybrid” wars are fought “also on the economic, financial and cyber fronts, where disinformation and campaigns that feed people’s fears are used to manipulate public opinion” so that increases in military expenditures are seen as the “only response” to an uncertain future. But all of this is only a “false realism”, an irresponsible Realpolitik that sows in people’s consciences and cultures a resignation to an unavoidable war and describes peace as a utopia (204-205). He notes that there is the possibility that some “may consider armed conflict as an effective way of diverting attention from domestic problems and a cynical tool for managing difficulties (208). The civilization of love: Christians, says Pope Leo, are called to respond to the culture of power by building “the civilization of love” and by choosing whether to feed the logic of force or safeguard peace. He recalls the memory of the saints, “righteous people and the oft-forgotten peacemakers, show us that grace does not magically eliminate conflict, but instead it inspires active resistance to evil and an astonishing creativity in doing good” (211). The Pope indicates five paths of responsibility, which include disarming words by speaking the truth; building peace in justice; adopting the perspective of victims by taking a stand, because there are conflicts in which “it is unjust to remain neutral”; cultivating “a healthy realism” that seeks practicable paths of peace through deeds, not only words. Importance of interreligious dialogue: Finally, relaunching dialogue by moving from a culture of power to a culture of negotiation is highlighted. And of decisive importance is “interreligious dialogue”, being bearers of a message of peace. He writes that “those who use the name of God to legitimize terrorism, violence or war betray his true nature, for to fight in the name of religion means attacking religion itself” (223). The diplomacy of the Holy See, he notes, “adopts the Gospel’s principle of mercy as a concrete criterion for political action.” And from this, the exhortation to prayer, since peace comes above all from God (227-228)." (https://www.vaticannews.va/en/pope/news/2026-05/pope-leo-xiv-encyclical-magnifica-humanitas-ai.html) more

Download (1)

Automated Disinformation and Malicious AI Swarms: Risks for Democracy and Development in Africa

Daniel Thilo Schroeder, Philipp M. Lutscher, Samba Dialimpa Badji, Stefan Brenner, Wenchao Dong, Tsehaye Haidemariam, Saheed Bidemi Ibrahim, Amanuel Tesfaye Kebede, Jonas R. Kunst, Johannes Langguth, Pauline Lemaire, Mulatu Alemayehu Moges, Lukasz Olejnik, Kristin Skare Orgeret, Gerald. Walulya

Online: arXiv, First (2026), 26 pp.

"Generative artificial intelligence is reshaping how information is produced, accessed, and circulated, while enabling disinformation campaigns of increasing scale and sophistication. There is currently no clear evidence that fully autonomous AI swarms conduct influence operations at scale, but thei ... more

External Link (215)

Building an Open and Responsible Voice Technology Ecosystem: Policy Recommendations for Digital Inclusion in India

Bhashini; Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ); FAIR Forward – Artificial Intelligence for All (2026), 57 pp.

"This report examines key barriers to building open and responsible speech systems in India—from data collection and model development to infrastructure and responsible practices. It proposes policy recommendations and governance mechanisms to support an innovative and equitable voice-technology e ... cosystem. Key Recommendations: A. TREATING FOUNDATIONAL DATASETS AS PUBLIC GOODS: Foundational datasets for speech technologies are large, reusable corpora of audio, text, and metadata. They are curated to support a wide array of downstream applications, including automatic speech recognition (ASR), text-to-speech (TTS), and speech translation. Making foundational datasets available as digital public goods addresses market failures in voice-technology ecosystems and promotes local economic innovation. The report identifies challenges arising from the linguistic diversity and nuance of Indian languages, infrastructural barriers, the absence of common data standards that create governance gaps, and unresolved intellectual property and privacy issues. Policy recommendations for building foundational language datasets include clarifying and revisiting existing laws to enable the use of publicly available material, ensuring sustainable investments supported by government and blended finance techniques, and instituting strong governance systems with shared standards, coordinated repositories, and independent quality assurance. B. BUILDING OPEN AND REPRESENTATIVE MODELS: India’s success in inclusive voice technologies depends on whether speech systems perform well across the country’s linguistic and social diversity. Today, limited openly available datasets, inadequate evaluation benchmarks, and uneven access to compute have resulted in models that perform inconsistently across languages, accents, and demographic groups, especially those from rural, low-resource, or marginalised communities. Strengthening India’s ecosystem requires further investment in data, evaluation, and infrastructure. Benchmarking should be built around open evaluation datasets and transparent leaderboards providing common baselines for developers, improving procurement standards, and helping assess performance across diverse languages and speaker groups. On the infrastructure side, platforms such as Bhashini, ULCA, and AI Kosh offer strong foundations although effective operation hinges on sustained governance, clear access protocols, and long-term funding models. C. INSTITUTIONALISING SUSTAINABLE OPEN-SOURCE INFRASTRUCTURE: Speech datasets place far greater demands on storage, bandwidth, and compute than text data, making the financing and governance of longterm hosting a central challenge for India’s voice-tech ecosystem. A fragmented licensing landscape adds further uncertainty: overlapping or incompatible terms across data, code, model weights, and evaluation sets impose substantial compliance burdens on small actors, while enforcement gaps allow misuse with little recourse. Sustaining open, equitable development requires treating dataset hosting as durable public digital infrastructure rather than grant-based, project-specific assets. By international standards, India’s emerging platforms, such as AI Kosh, provide a remarkable foundation. However, they require longterm funding, transparent governance, and clear access pathways for non-government contributors. Collaborative stewardship models, such as the Mozilla Data Collective, can help establish shared quality norms and consistent metadata conventions. D. STRENGTHENING RESPONSIBLE DEPLOYMENT: Deploying speech technologies responsibly requires more than high-performance models; it depends on safe systems, contextually appropriate use, and clear accountability. Existing data practices lack value-sharing mechanisms, leaving communities and researchers without recognition or benefitsharing, even as their contributions fuel commercial products. Risks of misuse, including voice cloning, phishing, and deepfake-driven misinformation, are rising, and unintended harms like linguistic exclusion, biased performance across accents and genders, and the erosion of regional language identities remain widespread. Addressing these gaps requires structural guardrails that embed fairness, transparency, and accountability into deployment workflows. Preventing misuse demands a combination of technical safeguards, stronger legal pathways, and widespread public literacy efforts to help users recognise risks and exercise their rights. The report argues that a strong ecosystem requires more than innovation funding. Building open-source foundations—including language datasets, standards, collection protocols and responsible AI frameworks—promotes demand-driven local innovation. It is therefore essential that the state plays an active, shaping role, much as it has in the development of digital public infrastructure. In the context of voice technology, this involves both investing in commercially viable languages and sustaining low-resource languages that are vital for inclusion but unlikely to attract private capital. Open-source assets can reduce costs for the public and private sector alike. However, they demand long-term planning and financing for hosting, maintenance, and updates. These assets can be supported through blended-finance models that pool public, philanthropic, and commercial resources. Emerging national initiatives, such as the proposed AI marketplaces, can further structure participation, transparency, and value-sharing across data, annotation, and deployable models." (Executive summary, pages 9-11) more

External Link (348)

Google Scholar (205)

WorldCat (336)

AI Companions Reduce Loneliness

Julian De Freitas, Zeliha Oğuz-Uğuralp, Ahmet Kaan Uğuralp, Stefano Puntoni

Journal of Consumer Research, volume 52, issue 6 (2026), pp. 1126–1148

"Chatbots are now able to engage in sophisticated conversations with consumers in the domain of relationships, providing a potential coping solution to widescale societal loneliness. Behavioral research provides little insight into whether these applications (apps) are effective at alleviating lonel ... more

Download (99)

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 (202)

Invisible Workers, Visible Harms: Perils and Precarities of AI Labour

Aapti Institute

Bonn: Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) (2026), 40 pp.

"Across South Asia, Southeast Asia, and Sub-Saharan Africa, data workers operate within fragmented and opaque supply chains characterised by informal employment arrangements, limited social protections, and a profound imbalance of power between workers, vendors, and the global technology firms that ... more

Google Scholar (67)

WorldCat (60)

Hallo, wie fühlst du dich heute? Wie künstliche Intelligenz unsere Beziehungen verändert

Christoph Koch

Zürich: Kein & Aber (2026), 272 pp.

"Christoph Koch hat ein Jahr lang mit Menschen gesprochen, die mit KI leben — mit denen, die sich verlieben, denen, die trauern, denen, die Chatbots als Therapie nutzen. Er hat mit KI-Forscherinnen und Psychologen geredet. Mit Anwälten und Geschädigten. Und mit den Maschinen selbst. Herausgekomm ... more

Download (36)

World Development Report 2026: The Promise of Artificial Intelligence
Deep Insights

Washington, DC: World Bank (2026), xxxi, 341 pp.

"This Report is the first comprehensive assessment of what artificial intelligence (AI) means for 6.8 billion people living in low- and middle-income countries. It argues that AI represents a historic opportunity for these countries, potentially enabling them to overcome development challenges that ... have defied solution for decades. Given the scarcity of information on AI’s use in developing countries, the Report sets the ball rolling on closing the gap. It uncovers new facts about how businesses and governments in developing countries are using AI—and identifies the main obstacles to greater and more productive use. AI is a general-purpose technology, or GPT. That puts it in rare company. Steam power. Electricity. Computers and the internet. Each transformed not just one sector, but entire economies, and did so continuously, over decades. Each also followed a pattern: slow and uncertain at first, but eventually reshaping how economies produce and how societies function. AI is following a similar pattern, but on a much faster timeline. It took about 80 years for the steam engine to reach lower-income countries, 40 years for electricity, and 20 years for the internet.1 By contrast, middle-income countries accounted for half of ChatGPT’s global traffic within six months of its launch.2 Rapid diffusion on such a scale has not been seen before. But history also teaches us that the countries that benefited most from previous GPTs were not always the ones that came up with the underlying technology. Those that benefited had built the necessary foundations and moved fast to develop applications on top of them. The countries that did not invest in the infrastructure, skills, and institutions that such technologies required to function benefited less and felt the most pain from disruption. This lesson is directly relevant to AI: The winning approach will take advantage of the new technologies, not seek refuge from them. In only a few years, AI has evolved from making predictions to generating content from text, audio, and video to now autonomously executing complex tasks. As it becomes more powerful and pervasive, its value as a tool for development depends on three factors. The first is capabilities. Unlike computers or the internet, AI can perform complex thinking tasks that guide decision-making and usually require human expertise, such as diagnosing diseases, forecasting weather, or advising farmers. These skills are often in short supply in developing countries, so AI has the potential to fill important gaps. The second is concentration. The most advanced AI models, the chips they rely on, and the data centers that run them are controlled by a small number of companies in just a few economies. This concentration creates dependency risks and affects whether the AI tools available will meet developing countries’ needs. However, this concentration also allows countries to customize AI models without having to spend billions of dollars creating the most advanced AI systems from scratch. The third is complements: the foundations that allow AI to be used safely and effectively. AI works best where reliable infrastructure, good education systems, and strong institutions are present. Developing countries often have shortages in all three. AI, moreover, needs to be customized to the local context because it learns from the data upon which it is trained. How quickly countries build these complements—and how well governments and businesses reorganize to use AI productively—will determine how much and how fast AI improves development. In that respect, AI is no different from GPTs that have come before. Ten years ago, World Development Report 2016: Digital Dividends called these prerequisites the analog complements to digital technologies. Today, AI requires both strong analog and digital complements to deliver its full benefits." (Overview, pages 1-2) more

External Link (199)

Advancing Open Source AI in India: Recommendations for Governments & Technology Developers

Aarushi Gupta, Anushka Jain, Urvashi Aneja, Shreeja Sen

Bonn; Delhi; Noida (IN): Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ); FAIR Forward – Artificial Intelligence for All; NASSCOM (2026), 80 pp.

"[...] this brief positions open source AI as a strategic policy option within India’s evolving AI ecosystem. It unpacks the critical importance of openness for India’s digital future, crystallising the key opportunities it offers to governments and developers alike. In doing so, it also explore ... more

Download (107)

Environmental Cost of AI's Energy Use: Carbon, Water and Land Footprints

Miriam Aczel, Sanaz Chamanara, Mir Matin, Aria Farsi, Tshilidzi Marwala, Kaveh Madani

Richmond Hill, Ontario (CA): United Nations University Institute for Water, Environment and Health (UNU-INWEH) (2026), 55 pp.

"[...] The study reveals some striking numbers. In 2025, data centers—the physical backbone of AI—consumed an estimated 448 TWh of electricity. If data centers were a country, that level of electricity use would rank it 11th globally. On current trajectories, data center electricity demand could ... roughly double to 945 TWh by 2030, nearly triple the combined annual electricity use of Pakistan, Bangladesh, and Nigeria, together home to more than 650 million people. Producing that much electricity would have a carbon footprint of 399 million tonnes CO2e, requiring 6.7 billion trees grown over 10 years to offset—roughly twice the number of trees in the United Kingdom. The associated water footprint of 9.3 trillion liters would be equivalent to the annual domestic water needs of all 1.3 billion residents of Sub-Saharan Africa. The land footprint associated with this electricity would exceed 14,500 km², nearly 10 times the size of Mexico City. AI is now one of the most significant drivers of that data center growth. In 2025, AI workloads alone accounted for around 20% of total data center electricity use, and if that share rises to 40% by 2030 as projected, its electricity demand could reach roughly 378 TWh—enough to meet the residential electricity needs of the entire population of Sub-Saharan Africa for over 2 years. AI’s environmental impacts are shaped not only by data center growth and electricity supply mixes, but also by the escalating cost of building ever-larger models. For example, GPT-3 training consumed an estimated 1.3 GWh of electricity over 34 days, while GPT-4 is estimated to have consumed 50 to 70 GWh over 100 days, roughly 40 to 55 times GPT-3. Yet training is only part of the picture as AI’s operational footprint is increasingly driven by inference. Once models are deployed, billions of everyday interactions account for the bulk of energy use, with inference estimated at roughly 80–90% of total energy consumption. [...] The report identifies priority implications for key stakeholders. Governments must integrate AI infrastructure into energy system planning, carbon accounting, water governance and land use permitting. This requires standardized environmental footprint reporting so impacts can be verified and compared across providers and jurisdictions. Industry and AI developers should treat model selection, default outputs, and routing decisions as footprint determinants, while also improving efficiency-by-design. Users and deploying organizations also shape impacts through volume, frequency, and modality choices. They should adopt “fit-for-purpose” use: selecting the lightest model and lowestenergy format that meets the task and limiting high-cost features when not needed. Data center operators and utilities should recognize siting and procurement as environmental footprint decisions, apply environmental impact screening and cumulative impact assessment. They should implement transparent mitigation and community safeguards where expansion occurs. Investors and financiers should treat electricity, carbon, water, and land footprints as material risks for AI infrastructure portfolios and use comparable footprint metrics in due diligence." (Executive summary, pages 12-13) more

Download (115)

Six Critical Lenses on AI-Generated Images

Catherine Bouko, Nataliia Laba (eds.)

CRC Press (2026), xvii, 176 pp.

"This book offers six critical, interdisciplinary approaches to AI-generated images. It invites the reader to (re-)imagine relationships between artefacts and media, authorship and agency, creativity and commodification, representation, and visual style. Across the chapters, the book examines three ... more

Download (87)

„Hi, how may I assist you?” Chatbots for Better Service Delivery. Practical Guide, Applications and Principles in Development Cooperation
Key Guides

Daniel Brumund, Prashant Pratap Singh, Lisa Werkmeister

Bonn; Eschborn: Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) (2026), 48 pp.

"The guide provides lessons-learnt for building AI chatbots and explores: motivations and requirements for developing a chatbot; functional types of AI chatbots suitable for different projects; content and knowledge base behind chatbots; technological architectures for AI chatbots; steering and man ... more

External Link (158)

The Essential Convergence: Global Compact on Extreme AI Risks

Sundeep Waslekar, Ilmas Futehally, Jayantika Kutty

Mumbai: Strategic Foresight Group (2026), 90 pp.

"Supply-Side Governance: Progress and Limitations: Most existing AI governance frameworks focus on the supply side of the technology ecosystem. These measures regulate developers, model providers, and infrastructure operators responsible for building advanced systems. Supply-side mechanisms include: ... pre-deployment safety testing and red-teaming; model evaluation and capability assessments; transparency and reporting obligations; incident response mechanisms; restrictions on certain high-risk uses. Several jurisdictions are experimenting with these tools in different combinations. China has implemented regulatory standards for generative AI systems; the European Union has adopted a risk based regulatory framework; Brazil and South Korea are developing legislative approaches; and the United States has introduced a mixture of executive measures and sectoral rules. These frameworks represent important progress. However, they face two structural limitations. First, supply-side regulation depends heavily on the cooperation of a small number of developer states. Second, the growing distribution of AI models, through open-weight releases, cloud APIs, and downstream applications, means that many systems reach global markets beyond the jurisdiction where they were originally developed. This reality limits the effectiveness of purely supply-side approaches. Demand-Side Governance: A New Dimension of AI Regulation: The paper introduces a second dimension of governance that has received far less attention: demand-side oversight. Demand-side governance shifts the locus of control from the point of creation to the point of deployment. Instead of relying solely on developer-state regulations, importing countries can establish safety requirements for any AI systems operating within their markets. These requirements may include risk assessments, certification standards, procurement rules, insurance obligations, and infrastructure safeguards. The central insight is that states controlling large consumer markets, digital infrastructure, financial systems, or public procurement channels possess significant leverage over global AI deployment, even if they do not develop frontier models themselves. This demand-side leverage can reshape incentives across the global AI ecosystem. Developers seeking access to major markets must comply with the safety conditions imposed by importing states. This approach represents a strategic shift in global AI governance: from a developer-centric model to a market-driven model of safety enforcement. Countries across different regions possess different forms of demand-side power. Brazil and South Africa represent large emerging markets; the United Arab Emirates plays a major role in AI infrastructure investment; South Korea combines technological capability with advanced regulatory institutions. When coordinated, these forms of leverage could help establish global safety norms. The Case for a Global Compact: Given the limitations of purely national approaches, the paper proposes the development of a Global Compact on Extreme AI Risks. The Compact would not aim to harmonise all aspects of AI regulation. Instead, it would focus narrowly on the most severe risk categories and establish shared mechanisms for prevention, monitoring, and response. Five institutional pillars form the backbone of the proposed Compact. The five pillars in the Global Compact should not be treated as isolated measures but designed as an interoperable system in which each pillar reinforces and informs the others through shared definitions, compatible standards, and coordinated implementation mechanisms. To address extreme risks effectively, the Compact should rest on a small set of guiding principles: a human-centric principle to ensure protection of humanity and human dignity as the primary objective; technological neutrality so that safeguards apply across methods and architectures rather than specific tools; interoperability to enable national and regional AI frameworks to exchange information, align risk classifications, and coordinate responses; and mutual cooperation to institutionalize cross-border collaboration in prevention, incident reporting, and crisis management. These can be complemented by principles of proportional risk governance, scientific integrity, and shared responsibility, creating a coherent normative base that allows diverse governance models to function together as a connected global safety architecture." (Executive summary, pages 5-6) more

External Link (209)

AlgorithmWatch’s guidelines to use generative AI responsibly

Oliver Marsh, Michele Loi

Berlin: Algorithm Watch (2026)

"Whether you use ChatGPT, Claude or Gemini, Copilot or Perplexity – generative AI poses massive problems: many results are inaccurate and politically problematic, the systems’ energy and water consumption is enormous. At the same time, they have become an integral part of everyday life. Algorith ... more

External Link (236)

Why Silicon Valley is turning to the Catholic Church: Priests and theologians want to shape the future of AI. Big Tech is listening

Elias Wachtel

The Atlantic, April 25 (2026)

"Silicon Valley and the Vatican may “seem like strange bedfellows”—but they each have something to gain from the other, @EliasWachtel argues. He examines why Big Tech is listening to priests and theologians who want to shape the future of AI." (Publisher description)

Download (130)

Societies at an AI Crossroads: Choices and Value Conflicts

Ignas Kalpokas, Julija Kalpokienė

London; New York: Routledge (2026), 116 pp.

"As contemporary societies undergo profound transformations driven by artificial intelligence and related technologies, a unilinear vision of inevitable ‘progress’ and development increasingly dominates public discourse. This book offers a counterpoint, challenging such deterministic narratives ... more

Download (160)

Reinventing Humanitarian Aid Procurement for the Age of AI

Giulio Coppi

Access Now (2026), 59 pp.

"Our research reveals that the organisational use of AI and algorithmic systems often happens through a process of unstructured integration rather than formal adoption. More precisely, cloud-based, algorithmic-enhanced functionalities and processes are making their way into humanitarian entitiesʼ t ... ech stack through updates and add-ons to preexisting products, slowly but inexorably seeping into most internal systems. A price-based procurement process designed around the static idea of digital asset purchase is inadequate to ensure adequate protection from hyperdynamic tech product development. This is especially evident looking at the growing power imbalances between private actors and nonprofits, and among nonprofits themselves. Digital technology is often heralded as key to democratise and equalise the playing field between humanitarian actors, by finally unlocking local actorsʼ access to resources traditionally reserved for international entities. But emerging trends seem to disprove such claims. The lack of adequate legal frameworks and safeguards, lagging investments in humanitarian funding for information and communication technology (ICT) resources and connectivity, and inadequate dedicated funding for digital programmes are thwarting these opportunities. More broadly, virtuous and responsible procurement practices are proving inadequate for protecting human rights alone, as tech companiesʼ commitments to human rights are replaced by defence-oriented pitches and the size of their human rights teams keeps shrinking. Preexisting digital divides are enhanced by several deep and subtle changes in traditional tech business models, resulting from the gradual cloudification of most common digital systems and the fast-paced, distributed nature of modern digital development. Even more worryingly, we are seeing new forms of digital divide well beyond the traditional gaps in digital capacity between local and small non-governmental organisations (NGOs) and international non-governmental organisations (INGOs) or UN entities. Cloud-based companiesʼ takeover of generic coordination platforms creates a constantly broadening chasm between the few organisations that are digitally mature, those who are barely retrofitting their processes to fit the new reality, and the broad majority forced to access the promised land of digital humanitarianism through the predatory gates of algorithmic and cloud servicesʼ free tech licensing. Our findings highlight how organisation-led digital development, building on open-source systems (despite some “openwashing” practices) and trusted external providers, remains the best option for risk reduction when it comes to algorithmic systems. However, in the short to medium term, this might not be financially and technically feasible for the majority of humanitarian actors, especially considering the catastrophic funding environment. To fill this gap, we propose a toolkit in an appendix to this report. This is a foundational governance model for transforming digital procurement from a transactional into a strategic business function, integrated across key organisational departments. By doing so, protection-mandated aid groups could move from running purchase-centred processes, to adopting a dynamic, proactive, and rights-based approach to tech management. This report also contains a set of recommendations for states and donors, the humanitarian community, tech companies, local aid actors and communities, research centres, and cyber experts." (Executive summary, pages 2-3) more

External Link (176)

Principled caution: How AI is used in public-interest media in Asia

James Gomez

Bonn: Deutsche Welle DW Akademie (2026), 52 pp.

"In conclusion, this research has provided a systematic understanding of AI adoption within public-interest media across Afghanistan, Bangladesh, Bhutan, Cambodia, Indonesia, Myanmar, Pakistan and Uzbekistan. By identifying concrete areas for strategic intervention, this research seeks to strengthen ... journalistic quality and prepare the sector for an AI-integrated future. The findings reveal that the AI adoption is currently taking place within public-interest media, but it is one of “principled caution” where adoption is careful, selective and highly varied. This evolution unfolds within a structurally precarious media environment where AI regulation is often absent or actively restrictive. While adoption remains cautious to protect journalistic credibility from risks such as hallucinations and biases, newsrooms are increasingly embracing AI as a powerful toolset to alleviate severe capacity and resource constraints. Currently, organizations primarily utilize AI for assistive, routine tasks such as translation, transcription, summarization, and early-stage content ideation and the final language editing stage rather than for broader organizational transformation. The research highlights deep capacity deficits in the governance framework, institutional training, and leadership. These gaps often prevent individual AI literacy from translating into consistent, organization-wide policy formulation and enforcement. This challenge is further exacerbated by the funding instability and resource constraints that characterize the region’s independent media, restricting the long-term investment necessary for ethical and responsible AI adoption. Despite these legal, political, technical, capacity, and operational constraints, localized and innovative AI workflows are emerging. While AI has not yet generated entirely new, revenue-generating business models, strategic investments in customized AI infrastructures, such as in-house dashboards and media monitoring systems, present a critical pathway toward long-term operational resilience. To support this trajectory, the report recommends coordinated action among media organizations, governments, technology companies, donors and media development funders to invest in long-term AI capacity, governance frameworks, and secure technological infrastructure." (Conclusion, page 39) more

Download (86)

Handbook on Data Protection and Privacy for Developers of Artificial Intelligence (AI) in India

Ajey Karthik, Aman Taneja, Sreenidhi Srinivasan, Vidushi Sinha

New Delhi: Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ); FAIR Forward – Artificial Intelligence for All (2026), 117 pp.

"This Handbook is designed as a practical guide for developers of AI systems, especially from early-stage startups, to navigate data protection obligations and ethical considerations, in a clear and actionable manner. Rather than serving as a comprehensive legal or technical manual, the Handbook co ... more

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

World Development Report 2026: The Promise of Artificial Intelligence Deep Insights

„Hi, how may I assist you?” Chatbots for Better Service Delivery. Practical Guide, Applications and Principles in Development Cooperation Key Guides

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

World Development Report 2026: The Promise of Artificial Intelligence
Deep Insights

„Hi, how may I assist you?” Chatbots for Better Service Delivery. Practical Guide, Applications and Principles in Development Cooperation
Key Guides