01 July 2026
CO26142 | Governing AI Beyond the Hype
SYNOPSIS
Global AI governance must move beyond industry hype and technological optimism by confronting the technology’s real socio-economic costs through evidence-based, people-centred policymaking. Drawing on the India AI Impact Summit and Pope Leo XIV’s Magnifica Humanitas, broader stakeholder participation and sustained accountability are needed to ensure AI serves the public good.
COMMENTARY
This year, leaders of AI firms have become more willing to justify the most pernicious externalities produced by AI technologies in the name of acceleration. In an interview at the India AI Impact Summit, OpenAI CEO Sam Altman not only rejected concerns about AI’s unprecedented energy consumption but also justified a data centre’s energy needs by comparing it to raising a human being and even argued that AI models have caught up to human cognition “on an energy efficiency basis.”
While Altman’s remarks, taken in isolation, may be construed as hyperbole, it is undeniable that many top AI firms have been grappling with the dilemma of pursuing cutting-edge innovations while managing negative externalities. In an open letter earlier in February 2026, a safety researcher resigned from Anthropic over what he described as “constant pressures to set aside what matters most”.
These recent actions are fuelling public anxieties over the technology. Satya Nadella, Microsoft’s CEO, once said that developers need to earn “social permission” to develop and deploy AI technologies for the industry to remain sustainable. This means the conversation around AI governance should be based on frank and evidence-based dialogue. Instead, a different kind of “AI bubble” is taking shape – not based on a gap between AI’s real value and value appraised by speculative investors, but rather between the effects the public views as risky and industry promises of positive upheaval to rationalise those risks.
Magnifica Humanitas and the Social Cost of AI
In his debut encyclical letter, Magnifica Humanitas, Pope Leo XIV cautions against what he calls “Babel syndrome”, or “dehumanisation”, by reducing the human experience to data and algorithms. He describes AI as a “valuable tool that requires vigilance” and calls for meaningful collaboration across sectors, including policymakers, scientists, businesses, civil society organisations, and religious institutions.
Through Magnifica Humanitas, Pope Leo helps unmask the true social cost of AI hype. One of the costs he highlights is labour disruptions. The Pope points out that AI automation is “transforming the very structure of work” and threatens to “erode workers’ sense of agency” as corporations continue to chase the hype. Indeed, he notes that an efficiency-only mentality will lead to significant disruptions and indiscriminate job losses.
Outside the Catholic Church, mainstream adoption of AI technologies is complicated by public anxieties. A Gallup study in the US found that nearly half of Gen Z employees now view AI’s risks as outweighing its benefits, even though other studies show that an overwhelming majority of US young adults regularly use AI chatbot services. This suggests that while public AI adoption is widespread, the industry has yet to gain the “social permission” Nadella argued was needed.
Reframing Global Action
What kind of global action is needed to meet this moment? While still useful, technical applications of ethical principles need to be accompanied by more decisive action to confront the real-world impacts that AI has already brought and will continue to generate.
The Delhi Declaration, issued at the conclusion of the India AI Impact Summit in February 2026, is a promising step forward. The declaration highlights key issues such as democratising access to AI models, enabling social empowerment, meaningful (re-)skilling initiatives, and environmental sustainability of models, which go beyond the technical backend of model development. Ninety countries endorsed the Declaration, a notable improvement on 2025’s AI Action Summit Declaration, which drew the support of around 60 countries.
As industry voices continue to gain influence in governance discourse, multilateralism will enable more holistic assessments of AI beyond the hype generated in Silicon Valley, and concerns from countries at different levels of digital maturity will also be given a chance to be heard. A strong multilateral approach also minimises the risk of global AI governance becoming a theatre of geopolitical competition driven by great power interests.
However, at this stage, many products of multilateral efforts like the Delhi Declaration are underpinned by voluntary frameworks. Turning them into sustained mechanisms for policy dialogue presents a different challenge. As Pope Leo suggests, outside the tech industry and policy circles, grassroots groups need to be folded in as key agents in the policy process. New multilateral and multistakeholder platforms may need to be formed to enable this. Still, policy input from a variety of sectors could deliver innovative policy solutions informed by real-world evidence.
This reframe also requires a rethinking of what industry leaders can do, what risks they are willing to bear, and what costs they think they can afford. Regulators used to be inhibited by concerns of prematurely constraining AI’s innovative promise. However, this thinking also entailed an implicit concession that in the short term, we could rely on the industry’s self-restraint and the threat of reputational costs for inaction to fend off some of these harmful risks.
Some industry leaders, at least on the surface, have been willing to engage. For example, Anthropic representatives were at the Vatican when Pope Leo unveiled Magnifica Humanitas. In April, OpenAI released a white paper titled Industrial Policy for the Intelligence Age, which outlined several steps to democratise access to AI, fortify security against its abuse, and safeguard workers’ welfare during the AI transition. Still, across the industry, and probably even within the boardrooms of Anthropic and OpenAI themselves, views on governance vary. Hence, continued engagement is necessary to ensure that these industry commitments can be sustained.
Conclusion
In their book The AI Con, Emily Bender and Alex Hanna argue that relying on absolutist hype or doom narratives toward AI will not help us understand how to reframe AI as a tool for wider social benefit. Indeed, AI’s social good lies somewhere in between. Policymakers should resist the temptation to accept, either passively or actively, the benefits and losses of AI use as articulated solely by tech companies themselves. It is only by critically and carefully assessing these perceived benefits and risks that AI can truly be harnessed for the social good.
About the Author
Jose Miguelito Enriquez is an Associate Research Fellow in the Centre for Multilateralism Studies at the S. Rajaratnam School of International Studies (RSIS) at Nanyang Technological University (NTU), Singapore.
SYNOPSIS
Global AI governance must move beyond industry hype and technological optimism by confronting the technology’s real socio-economic costs through evidence-based, people-centred policymaking. Drawing on the India AI Impact Summit and Pope Leo XIV’s Magnifica Humanitas, broader stakeholder participation and sustained accountability are needed to ensure AI serves the public good.
COMMENTARY
This year, leaders of AI firms have become more willing to justify the most pernicious externalities produced by AI technologies in the name of acceleration. In an interview at the India AI Impact Summit, OpenAI CEO Sam Altman not only rejected concerns about AI’s unprecedented energy consumption but also justified a data centre’s energy needs by comparing it to raising a human being and even argued that AI models have caught up to human cognition “on an energy efficiency basis.”
While Altman’s remarks, taken in isolation, may be construed as hyperbole, it is undeniable that many top AI firms have been grappling with the dilemma of pursuing cutting-edge innovations while managing negative externalities. In an open letter earlier in February 2026, a safety researcher resigned from Anthropic over what he described as “constant pressures to set aside what matters most”.
These recent actions are fuelling public anxieties over the technology. Satya Nadella, Microsoft’s CEO, once said that developers need to earn “social permission” to develop and deploy AI technologies for the industry to remain sustainable. This means the conversation around AI governance should be based on frank and evidence-based dialogue. Instead, a different kind of “AI bubble” is taking shape – not based on a gap between AI’s real value and value appraised by speculative investors, but rather between the effects the public views as risky and industry promises of positive upheaval to rationalise those risks.
Magnifica Humanitas and the Social Cost of AI
In his debut encyclical letter, Magnifica Humanitas, Pope Leo XIV cautions against what he calls “Babel syndrome”, or “dehumanisation”, by reducing the human experience to data and algorithms. He describes AI as a “valuable tool that requires vigilance” and calls for meaningful collaboration across sectors, including policymakers, scientists, businesses, civil society organisations, and religious institutions.
Through Magnifica Humanitas, Pope Leo helps unmask the true social cost of AI hype. One of the costs he highlights is labour disruptions. The Pope points out that AI automation is “transforming the very structure of work” and threatens to “erode workers’ sense of agency” as corporations continue to chase the hype. Indeed, he notes that an efficiency-only mentality will lead to significant disruptions and indiscriminate job losses.
Outside the Catholic Church, mainstream adoption of AI technologies is complicated by public anxieties. A Gallup study in the US found that nearly half of Gen Z employees now view AI’s risks as outweighing its benefits, even though other studies show that an overwhelming majority of US young adults regularly use AI chatbot services. This suggests that while public AI adoption is widespread, the industry has yet to gain the “social permission” Nadella argued was needed.
Reframing Global Action
What kind of global action is needed to meet this moment? While still useful, technical applications of ethical principles need to be accompanied by more decisive action to confront the real-world impacts that AI has already brought and will continue to generate.
The Delhi Declaration, issued at the conclusion of the India AI Impact Summit in February 2026, is a promising step forward. The declaration highlights key issues such as democratising access to AI models, enabling social empowerment, meaningful (re-)skilling initiatives, and environmental sustainability of models, which go beyond the technical backend of model development. Ninety countries endorsed the Declaration, a notable improvement on 2025’s AI Action Summit Declaration, which drew the support of around 60 countries.
As industry voices continue to gain influence in governance discourse, multilateralism will enable more holistic assessments of AI beyond the hype generated in Silicon Valley, and concerns from countries at different levels of digital maturity will also be given a chance to be heard. A strong multilateral approach also minimises the risk of global AI governance becoming a theatre of geopolitical competition driven by great power interests.
However, at this stage, many products of multilateral efforts like the Delhi Declaration are underpinned by voluntary frameworks. Turning them into sustained mechanisms for policy dialogue presents a different challenge. As Pope Leo suggests, outside the tech industry and policy circles, grassroots groups need to be folded in as key agents in the policy process. New multilateral and multistakeholder platforms may need to be formed to enable this. Still, policy input from a variety of sectors could deliver innovative policy solutions informed by real-world evidence.
This reframe also requires a rethinking of what industry leaders can do, what risks they are willing to bear, and what costs they think they can afford. Regulators used to be inhibited by concerns of prematurely constraining AI’s innovative promise. However, this thinking also entailed an implicit concession that in the short term, we could rely on the industry’s self-restraint and the threat of reputational costs for inaction to fend off some of these harmful risks.
Some industry leaders, at least on the surface, have been willing to engage. For example, Anthropic representatives were at the Vatican when Pope Leo unveiled Magnifica Humanitas. In April, OpenAI released a white paper titled Industrial Policy for the Intelligence Age, which outlined several steps to democratise access to AI, fortify security against its abuse, and safeguard workers’ welfare during the AI transition. Still, across the industry, and probably even within the boardrooms of Anthropic and OpenAI themselves, views on governance vary. Hence, continued engagement is necessary to ensure that these industry commitments can be sustained.
Conclusion
In their book The AI Con, Emily Bender and Alex Hanna argue that relying on absolutist hype or doom narratives toward AI will not help us understand how to reframe AI as a tool for wider social benefit. Indeed, AI’s social good lies somewhere in between. Policymakers should resist the temptation to accept, either passively or actively, the benefits and losses of AI use as articulated solely by tech companies themselves. It is only by critically and carefully assessing these perceived benefits and risks that AI can truly be harnessed for the social good.
About the Author
Jose Miguelito Enriquez is an Associate Research Fellow in the Centre for Multilateralism Studies at the S. Rajaratnam School of International Studies (RSIS) at Nanyang Technological University (NTU), Singapore.
