media development library

Filter

1404

Sort by:

Direction:

Navigating risks and rewards: How South African journalists use AI in the newsroom

Karen Allen, Herman Wasserman, Nande Mbekela

Johannesburg; Stellenbosch: Konrad-Adenauer-Stiftung (KAS), Media Programme Sub-Saharan Africa; Centre for Information Integrity in Africa (CINIA) (2026), 33 pp.

"This study demonstrates impressively how widely AI is already used in South African newsrooms – often, however, without systematic training, without editorial guidelines, and without tools adapted to African languages and contexts. The interviews make it clear that journalists stand between curio ... more

External Link (115)

Updated: Every African Country’s National Digital Health Strategy in 2026

Wayan Vota

ICTworks (2026)

"When I first compiled the list of every African country’s national eHealth strategy or digital health policy back in 2020, the dominant instrument across the continent was a five-year national eHealth strategy. They were written to a WHO Global Observatory template, often expired, frequently undi ... more

Download (105)

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 (78)

Targeted: How cuts to US foreign aid threaten freedom of expression and its defenders worldwide

London: Article 19 (2026), 112 pp.

"Between January and April 2025, the United States government implemented massive and immediate cuts to foreign aid spending. Our new report Targeted explores the global impact of those cuts on freedom of expression and the people working to defend it. The cuts have weakened civil society’s capac ... more

External Link (170)

AI in the media: Kenya and South Africa

Karen Allen, Herman Wasserman, Nande Mbekela

Bonn; Stellenbosch (ZA): Deutsche Welle DW Akademie; Centre for Information Integrity in Africa (CINIA) (2026), 27 pp.

"This study shows that AI remains largely an underexplored, and often misunderstood, topic for journalism in South Africa and Kenya. Stories can be framed in overly positive terms, such as the potential AI holds for development and social change, or negative terms, by highlighting the extractive nat ... ure of big tech firms or the potential of AI to foster disinformation. Either way, these reports frequently lack a more nuanced understanding of the impacts of AI in African contexts. Fear of the impact of AI on the media workforce may also explain why reporters say they often face resistance by editors when seeking to get AI and tech stories commissioned. AI articles tend to be siloed and reactive, framed as business, security or technology stories. Seldom is AI presented in terms of its multidimensional, layered impact on society. Furthermore, the geopolitical aspect of AI rarely gets interrogated in mainstream media. Journalists interviewed for the study admitted that many of their stories were reactive, and identified areas where synergies could be harnessed from other parts of the world. On issues of AI and social media regulation, for example, where precedents have been set by countries such as Australia, South African and Kenyan media could use such precedents to develop locally focused stories. AI reporting often adopts a framing promoted by technology vendors who derive interest from presenting a one-dimensional, optimistic view of AI. The unintended consequences often go unreported unless supported by an investigation, such as the CCTV camera surveillance reportage by Swart referenced above. Furthermore, the topic is often presented through a Global North lens, with little contextualisation or interrogation of these issues from an African point of view amid resource and access constraints. This occurs against a background of under-developed policy frameworks for AI on the continent generally, and in newsrooms in particular. Consequently the playing field can be tilted in favour of AI and tech vendors who fill the vacuum with hyperbole and hype about what the technology is capable of delivering. At the same time, more nuanced and specialised reporting on AI is constrained by the familiar obstacles of journalism, such as lack of resources, over-stretched newsrooms and story competition. To enhance coverage of AI in Africa, a multifaceted strategy is needed. While training and AI literacy programmes can deepen the knowledge needed by African journalists to investigate the topic more comprehensively, increased investment in newsrooms can provide the structural support to provide more context-specific coverage of AI on the continent. Crucially, training and education of African journalists should be done from within the African context, as solutions cannot be imposed from newsroom contexts in the Global North." (Conclusions, page 21) more

External Link (195)

Safer platforms, stronger participation: Mapping gender-based violence online in Syria

Bonn: Deutsche Welle DW Akademie (2026), 57 pp.

"Findings show that Gender-Based Violence Online( GBVO) in Syria is normalized and systemic, not merely isolated incidents, functioning as a mechanism of social control that reinforces patriarchal power and restricts civic participation. It takes overlapping forms, including sexual blackmail, harass ... more

External Link (149)

Countering Digital Scams: Stemming the Tide on an Urgent Development Challenge

Megan Roberts, Vitalii Zakhozhyi

UNDP (2026), 14 pp.

"Digital scams have become a global development challenge, inflicting costs of up to one trillion dollars annually. Beyond financial losses, scams inflict emotional harm and erode trust in digital systems, weakening confidence in governments, financial institutions, and online services. As countries ... more

External Link (191)

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 (167)

World Disasters Report 2026: Truth, Trust and Humanitarian Action in the Age of Harmful Information
Deep Insights

Charlotte Lindsey Curtet (ed.)

International Federation of Red Cross and Red Crescent Societies (ICRC) (2026), 372 pp.

"The World Disasters Report 2026 calls on governments, humanitarian actors, media, technology companies and communities, to recognise that the trustworthiness of information is a matter of life and death. Just as we plan for logistics, shelter and health care in emergencies, we must also plan for th ... e information environment. This requires investing in community engagement, prioritizing listening over speaking, building resilience against harmful narratives and consistently upholding humanity, neutrality, impartiality and independence in every interaction and message. In 2024, 32 Red Cross and Red Crescent volunteers and staff were killed while on duty; in 2025, a further 27 lives were lost. This is a painful reminder of the risks borne by those who put humanity first, often in the most dangerous circumstances. Their deaths underscore the urgent need to protect humanitarian staff and volunteers, and to safeguard the space they need to save lives. Anything that deepens mistrust, fuels hostility or turns the emblem of protection into a target must be confronted with urgency. Honouring their memory compels us to meet this crisis with urgency and resolve. Hope must also be part of our response. Harmful information thrives on fear and confusion, but hope – rooted in trust, solidarity and human dignity – can be just as contagious. Balancing threats with hope and fostering narratives of agency and possibility offers a vital counterforce. Around the world, communities respond to crises not only with resilience but also with creativity and compassion, often leading the way in finding solutions. By amplifying authentic stories and voices, we do more than counter falsehoods: we inspire action, strengthen trust and remind people that even in the darkest moments, there are paths forward. The IFRC and its member Red Cross and Red Crescent Societies have always worked at the heart of communities, earning trust through presence, transparency and respect. In today’s information crisis, this role matters more than ever. Our commitment is clear and unwavering: act with humanity and based on evidence, uphold trust and put people in need first – so that in moments of fear and uncertainty, the voices that carry furthest are those that heal, aid and protect. Each of us has a role to play. By staying informed, questioning, verifying and amplifying only what is trustworthy and constructive, we strengthen the information environment and ensure that hope speaks louder than harm." (Foreword) more

External Link (264)

The Local News Playbook: Creating Value for a Sustainable Future
Deep Insights

London; Miami: FT Strategies; Knight Foundation (2026), 105 pp.

"Drawing on global interviews with newsroom leaders, a cross-market survey of exemplary organisations, and FT Strategies’ work with more than 1,000 media businesses worldwide, the playbook identifies the core drivers that distinguish financially sustainable, community-rooted local news. It reframe ... more

Download (91)

Step by Step: How SAFE AI contributes to trustworthy AI

CDAC Network; Alan Turing Institute; Humanitarian AI Advisory (2026), 8 pp.

"Based on the National Institute of Standards and Technology (NIST) AI Risk Management Framework, the seven characteristics of trustworthy AI are: 1. Valid and Reliable: The AI system should perform as intended, delivering accurate results consistently across different situations. 2. Safe: The sy ... more

Download (105)

Generative AI and democracy: Study on advantages and risks of the use of generative artificial intelligence in public debate and democratic processes

Natali Helberger, Claes de Vreese, Gionata Bouche, Agustin Ferrari Braun, Max van Drunen, Simon Kruschinski, Nicolas Mattis, Sophie Morosoli, Laurens Naudts, Charis Papaevangelou, Theresa Seipp, Fabio Votta, Teresa Weikmann

Strasbourg: Council of Europe, Steering Committee on Democracy (CDDM) (2026), 77 pp.

"The key finding of this report is that it is not enough to make sure that the technology is safe and no risk to human rights and public values. GenAI’s impact and potential for the realisation of fundamental rights depends on the societal, organisation and regulatory context in which it operates, ... the possibility for institutions to make value-driven choices and regulators that actively promote public innovation where the interests of society as a whole outweigh purely economic interests of a concentrated handful of technology companies. This report has made the argument that GenAI and its implications for human rights and democracy must be seen in the context of the configurations of infrastructural and political power, and who is able and entitled to decide the course and character of innovation. [...] The report makes a number of recommendations to the Council of Europe: 1. Provide substantive guidance and standardisation: As a human rights organisation, the Council of Europe is particularly well prepared to develop more substantive guidance and identify, concretise and monitor risks to democracy, human rights and political equality, be that in the way GenAI is used in democratic practices like elections, the more structural implications for democratic institutions (such as the free press) as well as the democratic governance of AI. Next to a focus on risks, the Council of Europe can also usefully develop guidance how AI can advance democracy and human rights and develop a value-driven interpretation of “public interest innovation” to inform States, companies and standardisation processes. 2. Incentivise and enable public interest innovation: Next to the risked-based approach to AI that is also enshrined in the Council of Europe’s Framework Convention, the Council of Europe could develop guidance on policies and governance approaches that enable and incentivise public interest innovation. It could help identify the democratic and human rights implications of procurement choices, and develop narratives to prioritize value-oriented innovation, open-source datasets, decentralisation and interoperability above the promotion of scale and innovation driven by a concentrated number of technology companies. 3. Strengthen democratic resilience and ability to govern: The Council of Europe, and the CDDEM specifically, could have an important role in exploring mechanisms and regulatory solutions to map and remedy existing dependencies and identify mission-critical elements of a European digital ecosystem, for example through monitoring the use of GenAI in elections, stimulating the development of public interest information infrastructures and strengthening the ability of states to develop democratic and inclusive ways of governing GenAI. Next to digital autonomy, empowering public accountability actors like civil society, equality bodies, academics or investigative reporters to participate in the governance of the digital transformation are critical conditions for strengthening democratic resilience. Finally, democratic resilience also means enabling more meaningful citizen engagement, including of otherwise often marginalised groups, for example through transparency, literacy and meaningful participation in decision making." (Executive summary) more

Médias et démocratie: Enjeux et perspectives pour la philanthropie

Catherine Malaval, Gilles Marchand, Sabine Lenglet, Anne Cornilleau

Société de la Fondation de France, Observatoire Philanthropie (2026), 46 pp.

"À l’ère des algorithmes et de l’intelligence artificielle générative, l’enjeu démocratique n’est plus seulement l’accès à l’information, mais la capacité collective à identifier et reconnaître les acteurs de confiance qui en garantissent la production. Lorsque les images, les ... more

External Link (182)

The “Authoritarian Utopia”: The impact of funding cuts on independent media

Jos Midas Bartman, Evelien Wijkstra

Free Press Unlimited (2026), 23 pp.

"Recent reports by the Carnegie Endowment for International Peace (2026) and Article 19 (2026) on the impact of U.S. funding cuts, laid bare some of the most direct consequences of the U.S. funding cuts. They reported that the median grantee of their survey respondents lost about 35 to 40 percent of ... annual revenue for the year 2025. That number makes sense in light of the fact that the U.S. government cut at least $1.7 billion worth of freedom of expression programmes. This number was reported when some of the terminations were still unfolding. This is largely in line with our own findings, more than a year after the cuts. Out of the 57 surveyed partners, 37 percent reported budget cuts of 26-50 percent and 26 percent reported budget cuts of more than 50 percent Unfortunately, the abrupt and devastating level of the cuts have also led to media outlets and media support organisations ceasing operations. For this research we have not collected data on this but we are aware of several examples such as the closing of our Ukrainian partner, media outlet Zaborona, caused by the cessation of support of both the US and UK governments. Zaborona was an independent media outlet specialized in long-form, analytical reporting. The long-term political and geopolitical consequences of the U.S. governement funding cuts are still unfolding, but some of the medium-term effects are already visible. In the Philippines for example, disinformation has been able to flow more uncontrolled and narratives that support Chinese claims to the South China Sea have been diffusing more rapidly. Partners in Venezuela reported on how RT Academy provides training in several African countries, Argentina, Mexico and Cuba as an indirect way to approach people who can help amplify pro-Russian narratives in each of those countries, make friends and improve RT’s credibility. This report lays bare similar worrying but expected developments. Our survey shows unequivocally that where independent media decline, authoritarian interests fill the vacuum. Partners have reported on this dynamic across different continents. The boundary between domestic politics and international relations shows to be thin as well in this regard. In Pakistan for example, our partners report that China’s ability and assertiveness in trying to export its governance model have increased. Pakistani media organizations have been open about their recent collaboration with the Chinese embassy and the China–Pakistan Economic Corridor (CPEC). Authoritarian governments, such as Pakistan’s military elite, are unlikely to push back against these dynamics. [...] The same political actors who increasingly argue that international cooperation must be more interest driven should recognise that independent media serve precisely those interests. Where independent journalism collapses, corruption becomes harder to expose, authoritarian influence becomes easier to project, disinformation spreads with fewer constraints, and democratic allies become weaker. The vacuum is not neutral: it is filled by actors whose interests often run directly counter to those of the European Union. For Europe, the question is therefore not whether it can afford to support independent media globally. The question is whether it can afford not to. If the EU wants to remain a credible geopolitical actor, preserve its normative influence, strengthen democratic partners, and counter authoritarian interference, then support for independent media must be treated as strategic infrastructure: as essential to democratic resilience as election support, rule-of-law programming, anti-corruption work or cyber defence." (Executive summary, pages 4-5) more

Download (154)

Funding the Future of Local News: Lessons for the UK and beyond from journalism funds around the world
Deep Insights

Lucas Batt, Rhiannon Davies

Greater Community Media; Public Interest News Foundation (PINF) (2026), 124 pp.

"In this report we have documented ten case studies of visionary journalism support funds that we can learn from, drawing out insights and experiences from interviews with the people behind these funds. As we ourselves have been founders and leaders of pioneering local news organisations in the UK, ... more

External Link (169)

Tools for Reporting Online Violence Are Broken. Here’s How to Fix Them

Sherry Hakimi

Tech Policy Press, Aug 28 (2025)

"For several years now, we’ve seen leading social media platforms make dramatic cuts to their trust and safety teams, often in the name of ‘efficiency.’ Unfortunately, the results have been predictable: Harassment and online threats on those same platforms are more widespread than ever, and th ... more

External Link (289)

Declaration on the necessity of culture and media as a safeguard for our European democracies

Council of the European Union (2025), 4 pp.

"This declaration serves as a reminder that culture, cultural heritage and media policies must be a vital part of the European collaboration and European security in order to protect our democracies. We commit to upholding cultural diversity, artistic freedom and supporting media pluralism and freed ... more

Download (63)

Addressing Corrupt Uses of Artificial Intelligence

Berlin: Transparency International (2025), 14 pp.

"Transparency International is concerned specifically with the corrupt use of AI, defined as the intentional abuse of AI systems for the private benefit of entrusted powerholders. This is distinct from issues arising from unintended consequences and outcomes of the adoption of AI systems. In particu ... more

External Link (273)

Antiqua et nova: Note über das Verhältnis von künstlicher Intelligenz und menschlicher Intelligenz

Dikasterium für die Glaubenslehre, Dikasterium für die Kultur und die Bildung

Bonn: Sekretariat der Deutschen Bischofskonferenz (2025), 83 pp.

"Es ist daher von entscheidender Bedeutung, einzelne Anwendungen in bestimmten Kontexten kritisch zu bewerten, um festzustellen, ob sie die Würde und Berufung des Menschen und das Gemeinwohl fördern oder nicht. Wie bei vielen Technologien sind die Auswirkungen der verschiedenen KI-Anwendungen zu B ... more

Download (81)

Recommendation of the Council on Information Integrity

Paris: Organisation for Economic Cooperation and Development (OECD) (2025), 17 pp.

"The Recommendation on Information Integrity was adopted by the OECD Council on 17 December 2024 on the proposal of the Public Governance Committee (PGC). The Recommendation aims to establish a wide-ranging and applicable policy framework for Adherents to address threats posed by information manipul ... more

World Disasters Report 2026: Truth, Trust and Humanitarian Action in the Age of Harmful Information Deep Insights

The Local News Playbook: Creating Value for a Sustainable Future Deep Insights

Funding the Future of Local News: Lessons for the UK and beyond from journalism funds around the world Deep Insights

World Disasters Report 2026: Truth, Trust and Humanitarian Action in the Age of Harmful Information
Deep Insights

The Local News Playbook: Creating Value for a Sustainable Future
Deep Insights

Funding the Future of Local News: Lessons for the UK and beyond from journalism funds around the world
Deep Insights