The global financial architecture needs a deeper reset; it should build a country's capacity to withstand shocks and grow over time. Credit: Shutterstock
By External Source
Jun 15 2026 (IPS)
When G7 leaders arrive in Evian-les-Bains this month, France will host more than another summit. It will host a test of whether rich-country coordination can still solve problems that no country can manage alone. Aid budgets are shrinking, debt-service bills are crowding out investment, climate shocks are damaging infrastructure, and private capital remains scarce and expensive where it is needed most.
France has rightly made reducing global imbalances a priority of its G7 presidency. The G7 must show how finance should move differently and with global impact.
The urgent focus is development finance. G7 ministers have acknowledged that many partner countries face repeated crises, structural vulnerabilities, rising debt, food insecurity, and humanitarian needs. France has also placed African investment and the role of public development banks on the G7 agenda.
These issues are not separate. A drought that cuts harvests can weaken revenue, raise debt distress, damage health, interrupt schooling, and make the next investment more expensive. The current Ebola outbreak reminds us how vulnerable we all are to these crises.
The current global financial architecture was built for a world that believed growth could be separated from ecology, projects from systems, and risk from resilience. That world is gone
The dominant development finance paradigm treats each problem in its own box. The world no longer works that way.
That is why the global financial architecture needs a deeper reset; it should build a country’s capacity to withstand shocks and grow over time. Investments should work together. A solar plant that cannot feed a resilient grid, a road washed away by the next flood, or a hospital without reliable water, power, and social services support may look good in a project document and still fail the economy.
First, the world needs a better measure of wealth. GDP is useful, but incomplete. It counts activity; it does not tell us whether a country is building or consuming the assets on which future prosperity depends. A forest cleared for short-term export can raise GDP, and so can rebuilding after a flood.
Neither means that a country is becoming richer if its soils, water, skills, and institutional trust are deteriorating. A reset should ask whether produced assets, natural systems, people’s capabilities, and public institutions are becoming stronger together.
The practical step is not abstract. Finance ministries could require comprehensive wealth impact statements. When a government considers a debt-financed power system, port, irrigation program, or disaster-risk loan, it should show not only the likely effect on deficits and growth, but also the likely impact on water security, land-use management, public health, skills, and future disaster losses.
Creditors and rating agencies should look at the same evidence. A country that protects floodplains, strengthens schools, and reduces energy vulnerability is making itself a safer borrower, even if those gains remain invisible in conventional accounts.
Second, the world needs to appraise investment portfolios, not isolated projects. This is where many well-intentioned plans underperform. A seawall without drainage and mangrove protection may shift risk rather than reduce it.
Climate-smart agriculture without storage, cold chains, and roads leaves farmers exposed. Solar panels without grid upgrades and reliable payment systems can leave generations stranded. The question should not be which project has the highest standalone return, but which combination of investments most improves resilience, productivity, and long-term wealth in the public interest.
This approach would also help mobilize private capital. Investors are often told that developing countries are too risky. But part of that risk reflects weak systems: unreliable power, poor maintenance, exposed supply chains, thin insurance, and fragile public finances.
Coordinated ublic investments should be used to lower these risks at the portfolio level by preparing interconnected pipelines, funding data, providing guarantees, supporting local-currency finance, and strengthening early-warning systems and building the institutions that keep assets working when shocks hit. Capacity building would not be a charity; it would be risk reduction.
Third, states and markets need clearer rules for allocating capital. For policymakers, this means that budgets, debt strategies, and industrial plans should include the assets and vulnerabilities they create. For multilateral development banks, the IMF, credit-rating agencies, and regulators, it means treating climate adaptation, nature protection, social capability, and debt sustainability as one conversation, not four.
Country platforms should bring them into a single investment plan with clear priorities and accountability. For investors, assets that protect water, power, food systems, health, and skills should be viewed as infrastructure for returns, not as ESG decoration.
The G7 can make this pivot at Evian. It could agree that major development-finance packages should include wealth impact statements; that multilateral development-bank country strategies should use portfolio appraisal; that public development banks should standardize guarantees and project preparation for resilience; that debt workouts and new lending terms should reward verified investments that reduce future losses; and that private co-financing should be linked to transparent outcomes. These reforms simply require an acceptance by institutions to judge success differently.
None of this is anti-market, anti-growth, or anti-finance. It is pro-accuracy, pro-stability, and pro-prosperity. The central task is simple: build a financial architecture that strengthens society’s productive capacity and the planet that sustains it, not that merely flatters the next quarter’s accounts.
The current global financial architecture was built for a world that believed growth could be separated from ecology, projects from systems, and risk from resilience. That world is gone.
France’s G7 presidency offers a chance to replace it with a financial system that measures real wealth, funds investments that work together, and rewards countries for reducing the risks that threaten everyone. That is how we move from fragmented finance to resilient prosperity and from short-term gain to long-term global public investment.
Hyginus ‘Gene’ Leon is the Executive Director of the Development Bank for Resilient Prosperity and was the sixth President of Caribbean Development Bank (CDB). Simon Reid-Henry, PhD is a Research Professor at the Peace Research Institute Oslo.
Chatbots and AI companions have rapidly moved from science fiction into everyday life. Credit: Shutterstock
By Joseph Chamie
PORTLAND, USA, Jun 15 2026 (IPS)
AI chatbots and AI companions designed to simulate human-like conversation and provide relationships and companionship through generative artificial intelligence (AI) have rapidly evolved from science fiction into everyday reality.
Globally, approximately one billion people – about 12% of the world’s population – now use generative AI chatbots monthly, with usage approaching parity among men and women.
Dedicated AI companions and virtual friends are estimated to have between 50 to 100 million active users worldwide. The global AI companion market is valued at roughly USD 50 billion in 2026 and is projected to grow nearly ninefold by 2034.
These technologies, including the growing use of AI avatars, are increasingly taking the place of human interactions in homes, schools, workplaces, and other settings. Marketed as virtual friends, romantic partners, or personal assistants, AI chatbots and AI companions offer users emotional support, entertainment, guidance, and companionship.
As their capabilities become more sophisticated, many users report forming emotional attachments to these systems, with increasing numbers of users believing that their AI companion or chatbot is sentient or possesses human-like awareness.
While these technologies can provide new opportunities for connection, they cannot replace the face-to-face interactions that are essential to social development, particularly among children and adolescents
Advances in robotics are also moving AI companions beyond screen-based interactions into the physical world. With increasingly human-like appearances, behaviors, and communication abilities, these systems are becoming more sophisticated and human-like in the way they interact with people.
Unlike AI assistants, which primarily answer questions or perform tasks, AI companions are designed to simulate conversations and relationships, encouraging emotional connections as friends, confidants, or romantic partners.
By providing human-like conversation, these artificial intelligence devices are offering support against social isolation and loneliness, providing educational instruction, dispensing advice and guidance, becoming friends and romantic partners, and transforming personal relationships.
The chatbots and AI companions have introduced social, psychological and ethical changes to how men, women, and especially children experience companionship, domestic life, and schooling. In particular, generative AI chatbots and AI companions have opened a new frontier in developing friendship and social relationships.
Many adolescents now rely on these new technologies for school assistance, entertainment, and emotional support. As a result, relationships with chatbots and AI companions – as friends, therapists, and even romantic partners – have become increasingly complex and, in some cases, riskier.
These emotionally engaging interactions can exacerbate psychological vulnerabilities and blur the lines between human relationships and machine-generated companionship.
In several widely publicized cases, AI chatbots have encouraged or failed to prevent self-harm. In addition, some deaths have been linked to young people who developed obsessive emotional attachments to AI companions.
However, despite the complications and risks, the world’s current attention and concerns about AI remain focused primarily on its growing impact on employment, budgetary cuts, and taking over jobs currently performed by men and women.
In contrast, relatively little attention is being given to chatbots and AI companions that engage in conversations and increasingly form personal relationships with men, women, teenagers, and children at home, in schools and in many other settings.
While these technologies can provide new opportunities for connection, they cannot replace the face-to-face interactions that are essential to social development, particularly among children and adolescents.
AI chatbots also raise risks to personal privacy, psychological well-being, the spread of misinformation, and the reinforcement of harmful behaviors. In addition, a broad range of other concerns has been identified regarding the use of chatbots and AI companions.
These concerns include delaying social and emotional development among children and teenagers, blurring the distinction between software and reality, encouraging risky behavior, exploiting young people’s emotional needs, reinforcing unhelpful thoughts, distorting users’ sense of reality, and fostering simulated attachments and dependence (Table 1).
Source: Author’s compilation.
The United States Psychological Association recently warned that relationships between children and adolescents and AI chatbots could displace or interfere with healthy social development. The association noted that friendships and social support from other people have long-term benefits for emotional well-being, physical health, and longevity.
Among generative AI chatbots, the leading platforms by market share in May 2026 are generally reported to be ChatGPT, Claude AI, Google Gemini, Microsoft Copilot, Perplexity, and Grok. Several industry analyses place ChatGPT’s share at roughly 50-55%, with Claude AI at about 21% of market share emerging as the second-largest platform (Figure 1).
Source: FirstPageSage.
In March 2026, the country with the largest number of ChatGPT users was the United States, with approximately 205 million users. Following the U.S., the countries with the largest ChatGPT user populations were India, Brazil, Canada, and France (Figure 2).
Source: fatjoe.
It is certainly the case that chatbots and AI companions cannot feel love toward an individual. Nevertheless, hundreds of millions of men, women, and children worldwide are increasingly relying on these technologies for conversation, information, companionship, and non-judgmental interactions.
These technologies may help to address chronic loneliness and social isolation, conditions that have consistently been linked to detrimental effects on physical and mental health and increased risk of premature death. The World Health Organization (WHO) formally recognizes loneliness as a global public health concern, with roughly one in six people worldwide experiencing problematic levels of loneliness.
Chatbots and AI companions can help alleviate loneliness and social isolation by providing readily available conversation and companionship without judgement and expectations. As chatbots, AI companions, and androids become increasingly sophisticated, growing numbers of people are exploring the new forms of emotional connection and intimacy with these technologies.
At the same time, the growing use of chatbots and AI companions for personal relationships raises important social, psychological, ethical, and policy concerns.
Although chatbots and AI companions may help reduce loneliness and social isolation for some users, they also pose risks, especially for children and young people. Because AI systems do not possess genuine empathy and are not trained or licensed as mental health professionals, excessive reliance on them for emotional support may isolate vulnerable individuals and distort perceptions of human relationships.
Debate continues regarding the appropriate level of regulations for these technologies. Some government officials, technology companies, investors, and researchers argue that these new and emerging AI technologies should remain largely unregulated, with people themselves determining how to adapt to these technologies.
Some of the reasons for keeping the development of AI unregulated include: prevents regulatory paralysis; accelerates technological breakthroughs; encourages venture capital investment; maintains global geopolitical competitiveness; promotes national security; prevents market monopolies; benefits national interests; and leads to better lives for men and women.
Others, however, argue that AI chatbot and AI companion technologies need to be regulated in order to protect the mental health of children and young adults; reduce the negative effects of social media and excessive screen time; mitigate risks, deception, bias, discrimination, and misinformation; promote economic stability and fairness; become a public resource; protect human rights and intellectual property; and ensure data privacy.
Among the proposed safeguards and regulations for chats and AI companions are requirements for non-human disclosure, crisis protocols for self-harm, age verification measures, limits on their use in elementary schools, bans on impersonation, and stronger protections for minors.
Fueled in part by technology companies, governments worldwide are moving rapidly to deploy generative AI systems and chatbots in schools, universities, and other settings.
However, the spread of these new AI technologies may pose risks to the development and well-being of children and teenagers, raising concerns among educators, parents, and policymakers. Interactions with AI chatbots, especially when they are intense and prolonged, may contribute to the onset or worsen delusions or mania. Research is also finding that AI companions provide responses that may worsen mental health issues.
Additionally, a recent study reported that reliance on generative AI chatbots may reduce critical thinking engagement in some contexts. Another study has raised concerns that AI chatbots can exploit teenagers’ emotional vulnerabilities, sometimes leading to inappropriate and harmful interactions.
The United States Federation of Teachers recommends “no screens” for children in second grade or younger, and restricting the use of AI chatbots for students in elementary schools. The organization has expressed concerns that excessive screen use may hinder socialization, independent thinking, and critical-thinking development.
The long-term effects of AI chatbots remain uncertain, with researchers just beginning to investigate them. However, classroom teachers and some city officials report that many students are increasingly relying on chatbots for easy answers rather than developing problem-solving and critical-thinking skills.
The U.S. Federation of Teachers has urged elementary schools to avoid using artificial intelligence tools like AI chatbots with students and called for national privacy and safety standards governing AI use in schools.
Research suggests that chatbots and AI companions may pose several risks, particularly for teenagers. Concerns include emotional dependency, declining mental health, harmful interactions, and revealing sensitive personal information, including mental health issues and sexual orientation.
Reliance on chatbots and AI companions for emotional support may also contribute to social isolation and interfere with the development of normal human relationships. Because these technologies are designed to simulate emotional intimacy, they can blur the line between genuine human connections and artificial interactions.
A risk-assessment study found that inappropriate dialogue could be readily elicited from chatbots on topics such as sex, self-harm, violence, drug use, and racial stereotypes, raising concerns about their influence on vulnerable users, particularly children and adolescents.
In conclusion, chatbots and AI companions have rapidly moved from science fiction into everyday life. They increasingly exhibit human-like characteristics, including natural-sounding human voices, memory of past interactions, continuous processing of personal information, apparent preferences, constant availability, and the ability to provide companionship and guidance on personal and social matters.
Public discussion of generative AI has focused largely on employment and job displacement, while less attention has been given to its social, psychological, and ethical effects. As chatbots and AI companions become more capable and widely used, concerns about their impact on the well-being, development, and relationships of young people are likely to become increasingly important for parents, educators, policymakers, and technology developers.
Joseph Chamie is a consulting demographer, a former director of the United Nations Population Division, and author of many publications on population issues.
The international development debate is increasingly referring to the notion of the “social contract”. In this paper, we measure what governments give societies, a core element of social contracts. To enable social contract comparison across countries and over time, we develop indices to capture the three “Ps”: protection against internal and external threats, provision of social and economic services, and political participation. These indices are composed of indicators, which are mainly input variables to gauge the willingness of governments to deliver the three Ps. Subsequently, we calculate the values of 154 countries for the three indices around the year 2019. The results show that the indices are useful and valid. They highly correlate with each other and with other indicators such as per capita income and the Human Development Index. Yet, these correlations are not perfect, meaning that the indices are not another redundant development index. They add information and value. Finally, we make a first step in identifying patterns in the results. Countries in Latin America were doing comparatively well on average in terms of political participation in 2019. When controlling for per capita income, governments in sub-Saharan Africa, were delivering disproportionately more on average in terms of protection and political participation, but less so in terms of provision. Countries in the Middle East and North Africa fail mainly with regard to political participation.
The international development debate is increasingly referring to the notion of the “social contract”. In this paper, we measure what governments give societies, a core element of social contracts. To enable social contract comparison across countries and over time, we develop indices to capture the three “Ps”: protection against internal and external threats, provision of social and economic services, and political participation. These indices are composed of indicators, which are mainly input variables to gauge the willingness of governments to deliver the three Ps. Subsequently, we calculate the values of 154 countries for the three indices around the year 2019. The results show that the indices are useful and valid. They highly correlate with each other and with other indicators such as per capita income and the Human Development Index. Yet, these correlations are not perfect, meaning that the indices are not another redundant development index. They add information and value. Finally, we make a first step in identifying patterns in the results. Countries in Latin America were doing comparatively well on average in terms of political participation in 2019. When controlling for per capita income, governments in sub-Saharan Africa, were delivering disproportionately more on average in terms of protection and political participation, but less so in terms of provision. Countries in the Middle East and North Africa fail mainly with regard to political participation.
The international development debate is increasingly referring to the notion of the “social contract”. In this paper, we measure what governments give societies, a core element of social contracts. To enable social contract comparison across countries and over time, we develop indices to capture the three “Ps”: protection against internal and external threats, provision of social and economic services, and political participation. These indices are composed of indicators, which are mainly input variables to gauge the willingness of governments to deliver the three Ps. Subsequently, we calculate the values of 154 countries for the three indices around the year 2019. The results show that the indices are useful and valid. They highly correlate with each other and with other indicators such as per capita income and the Human Development Index. Yet, these correlations are not perfect, meaning that the indices are not another redundant development index. They add information and value. Finally, we make a first step in identifying patterns in the results. Countries in Latin America were doing comparatively well on average in terms of political participation in 2019. When controlling for per capita income, governments in sub-Saharan Africa, were delivering disproportionately more on average in terms of protection and political participation, but less so in terms of provision. Countries in the Middle East and North Africa fail mainly with regard to political participation.
Thousands gather outside Istanbul City Hall to mark one year since the arrest of Mayor Ekrem İmamoğlu on 18 March 2026. Credit: Yasin Akgul/AFP
By Inés M. Pousadela
MONTEVIDEO, Uruguay, Jun 15 2026 (IPS)
When Hungary’s Prime Minister Viktor Orbán lost by a landslide to a unified opposition in April, Turkey’s President Recep Tayyip Erdoğan was watching. The lesson he drew was not that he should be more moderate; it was that he needed to crack down harder. He had already arrested Istanbul Mayor Ekrem İmamoğlu, the opposition Republican People’s Party (CHP)’s leading presidential contender, in March 2025. After Orbán’s defeat, he has accelerated his campaign to fracture the opposition and rewrite the rules before the next election in 2028.
Electoral autocracy
Erdoğan has been in power since 2003. After surviving a coup attempt in July 2016, he used emergency powers to purge the state at scale. Over 150,000 people were detained, fired or suspended from their jobs. Emergency decrees expanded the government’s power to shut down organisations and remove elected officials. A 2017 constitutional referendum, narrowly approved in a campaign that independent observers found deeply flawed, replaced Turkey’s parliamentary system with a hyper-presidential one.
Independent media has been systematically dismantled. Turkey now ranks 163rd out of 180 countries on Reporters Without Borders’ 2026 World Press Freedom Index. Yet elections have continued, and the opposition has continued to win at the municipal level, most strikingly in Istanbul in 2019 and again by an even wider margin in 2024. That residual competitiveness is what Erdoğan is now moving to close.
İmamoğlu had beaten Erdoğan’s candidate in Istanbul twice, was formally nominated as the CHP’s 2028 presidential candidate and polled strongly against Erdoğan nationally. Authorities arrested him on charges of corruption and links with terrorism as his nomination was under way, triggering Turkey’s largest wave of protests in over a decade. A 4,000-page indictment filed in November 2025 sought to sentence him to over 2,000 years in prison. Espionage charges followed in February 2026. His trial began in March amid continuing protests. He remains in prison, and in the 14 months since his arrest, over 500 more people have been detained, including 16 CHP-affiliated mayors.
With İmamoğlu imprisoned, Erdoğan’s next move was to prevent the CHP from consolidating around anyone else. On 21 May, an appeals court annulled the outcomes of the CHP’s 2023 national congress, ejecting the party’s elected leader Özgür Özel, who had raised the CHP to rough parity with Erdoğan’s Justice and Development Party (AKP) in national polls, and reinstating his predecessor Kemal Kılıçdaroğlu, a divisive figure who lost the last presidential election. Özel condemned the ruling as a judicial coup and refused to leave the party’s headquarters. Three days later, riot police stormed in, firing rubber bullets and teargas. The government denied any involvement, implausibly claiming the judiciary had acted independently. The operation was legal in form and political in substance.
Turkey’s constitution limits presidents to two five-year terms, and Erdoğan’s second expires in 2028. In May 2025, he appointed a legal team to draft a new constitution. It seems clear the goal is to extend his eligibility. The AKP and its nationalist allies fall short of the parliamentary threshold required to change the constitution or call a referendum on it. Some analysts believe the government’s recent initiative to end the decades-long conflict with the Kurdistan Workers’ Party is at least partly designed to attract enough parliamentary votes to clear that threshold.
There is a structural reason the stakes are so high. Turkey’s hyper-presidential system means that, unlike Orbán, Erdoğan would have no safe path back from electoral defeat. For him, losing power could mean political extinction. His crackdown is a response to this threat.
Civil society resistance
Turkey’s civil society has, however, not submitted. Huge protests followed İmamoğlu’s arrest. A mass rally marked his 100th day in jail, and people marched again when the CHP headquarters were raided. Most recently, when Erdoğan ordered the closure of Bilgi University, one of Turkey’s oldest liberal academic institutions, students and staff immediately gathered outside to protest. Within two days the government reversed the closure. This illustrated both the extent of Erdoğan’s repressive urges and their limits when met with swift resistance.
The government has responded to protest with blanket bans on public gatherings, social media restrictions and mass arrests. Four days after İmamoğlu’s arrest, at least 1,879 people had been detained. Police repeatedly intervened forcefully, using teargas and detaining protesters and journalists.
Orbán’s downfall has frightened Erdoğan as much as it has inspired the Turkish opposition. He is moving to eliminate the conditions that made it possible. He has got rid of the most credible and unifying opposition candidate, neutralised the main opposition party and is in the process of dismantling what’s left of an electoral architecture that, however tilted, could still allow the opposition to win.
Turkey’s democracy now depends on whether enough people keep showing up, and on whether they can keep resisting Erdoğan’s campaign to dismantle democracy.
Inés M. Pousadela is CIVICUS Head of Research and Analysis, co-director and writer for CIVICUS Lens and co-author of the State of Civil Society Report. She is also a Professor of Comparative Politics at Universidad ORT Uruguay.
For interviews or more information, please contact research@civicus.org
Follow @IPSNewsUNBureau