Towards a Collective Ownership of AI

Artificial intelligence has ceased to be merely a disruptive innovation, or even an industrial sector. It is a transformative force, bringing about daily upheavals in the world of work, in value chains, in educationAccountability within the public education system is key to improving outcomes and attainment, and accountability is nearly impossible without transparent policies and opportunities for participation ..., but also in our imaginations and our relationship with the world.

Little by little, it is creeping into the very fabric of our lives.

The problem? This revolution is essentially driven by a handful of companies, investors and researchers. The efforts of governments and the international community have undoubtedly succeeded in forging a consensus on the broad principles that should apply to the development of this new civilisational order, but they have done almost nothing to alter Silicon Valley’s trajectory.

Global discussions of “AI safety” have often been framed as avoiding the unknown existential risks of tomorrow, while real-world harms of today often go unaddressed. And Europe’s commendable efforts to “catch up” or establish regulatory frameworks such as the GDPR, the Digital Services Act, the Data Act, AI Act also miss the most essential point: this revolution must be democratic. It is up to us, collectively and following an informed debate, to decide on a desirable future.

Let us be clear: to date, there is no structured international governance of AI. Even though groups like the G7 have set principles, there is an opportunity for global leaders to go a step further. The Global Digital Compact has laid the groundwork. National and regional regulations are emerging, starting with the European AI Act that came into force in 2025. But no single global framework exists to guarantee that this technology will serve human dignity, social justiceTo address barriers that prevent citizens from having their justice needs met, OGP participating governments are working to expand transparency, accountability, and inclusion into all systems of justi... and democratic balance. The democratic legitimacy of AI can only come from ongoing, open deliberation involving those who are affected by it.

As the G7 met last week to discuss the future of AI, members had a unique opportunity to build on past efforts and set a path for open deliberation to involve those affected by unregulated digital technologies.

A democratic crisis before a technological one

However, without even mentioning the concentration of the market around a narrow oligopoly, and the risk of capturing entire sectors of the economy, without revisiting the need to consider and support the radical transformations of the world of work, and without dwelling on the environmental cost of the paths currently being taken – all without any real debate – it is clear that the first extremely worrying democratic accidents have already occurred.

The Robodebt scandal in Australia, the A-level marking algorithm in the UK, the fraud detection system in the Netherlands: these cases starkly illustrate the human cost of systems deployed without safeguards, without independent oversight, and without effective recourse. The biases in facial recognition and predictive policing systems have reinforced existing discrimination and eroded trust in institutions.

Worse still, perhaps, at a time when our representative democracies are weakened by a crisis of institutions and representation – a crisis that did not begin with AI – the growing integration of AI into decision-making processes, which silently imposes its algorithmic rationality, dilutes human accountability, and masks the effects of power, promises to further exacerbate a crisis that is undermining open democracies.

AI and the challenge of relational equality

AI is amplifying a tension that digital technology had already brought to light. On the one hand, widespread access to information and the capacity for distributed analysis promise an unprecedented level of horizontal power: every citizen can verify, challenge and propose. On the other, this very same horizontal structure undermines traditional intermediaries — political parties, trade unions, the media, public administrations — without which responsibility and accountability become impossible.

AI can be used to strengthen democratic processes: simplifying access to public services, detecting fraud, informing complex decisions, and making institutions more responsive. But it can just as easily undermine them: obscuring decisions, automating discrimination, and concentrating power in the hands of those who control the models.

The issue is political, and it is simple: do we want AI that strengthens citizens’ ability to understand, deliberate and decide? Or AI that strips them of that power? Will it be used to empower us? Or to turn us into passive, docile vessels?

A French legacy

Almost ten years ago, in December 2016, Paris hosted the Open Government Partnership (OGP) Global Summit. At the time, France chaired this unprecedented alliance — consisting of representatives from more than 70 countries and hundreds of civil society organisations, united by a shared conviction: transparencyAccording to OGP’s Articles of Governance, transparency occurs when “government-held information (including on activities and decisions) is open, comprehensive, timely, freely available to the pub... More, citizen participationAccording to OGP’s Articles of Governance, citizen participation occurs when “governments seek to mobilize citizens to engage in public debate, provide input, and make contributions that lead to m... More and government accountability are not democratic luxuries, but essential for the survival of our institutions.

In the history of the relationship between democracy and technology, these ten years are an eternity. We have witnessed the weakening of the order based on international law, attacks – whether state-sponsored or not – on the rule of law, and the capture and distortion of the public sphere by social media.

Today, the rapid emergence of artificial intelligence raises fears of even worse threats: an unprecedented concentration of power without checks and balances, unregulated uses of AI in the information sphere, and the proliferation of cyber intrusion tools.

France has consistently placed technopolitical issues – and questions of digital governance – on its international agenda. In addition to its ongoing commitment to multilateralism and its participation in multi-stakeholder dialogue forums, embodied each year in Paris by the Paris Peace Forum, it has also demonstrated a consistent commitment to seeking open and shared governance of various aspects of the digital sphere. France’s leadership role has led to the Paris Call for Trust and Security in Cyberspace in 2018, the Christchurch Call against terrorist content in 2019, the Global Partnership on Artificial Intelligence in 2020, and the Pall Mall Process on cyber intrusion tools in 2024. In February 2025, at the Summit for Action on AI, France once again brought together the international community, seeking to build a multi-stakeholder consensus where divisions are deepening.

What links all these initiatives? They are rooted in the same core conviction. Digital revolutions, like all industrial revolutions, are also political revolutions. They call for a democratic response, which can only be developed through close multi-stakeholder dialogue involving governments, civil society, researchers and businesses. And broad principles are not enough: infrastructure, institutions and, above all, concrete cooperation are an integral part of the solution.

The work of activists ten years ago foreshadowed the excesses we are now witnessing. It proclaimed that democracy was not a guaranteed right, and that it needed to be protected. Ten years on, it may be necessary to state that, conversely, the destruction of social bonds, solidarity and human dignity is not inevitable either.

The quest for open governance

Beyond the broad principles – which are so difficult to put into practice – the world of multi-stakeholder collectives and open democracy is seeing a proliferation of local initiatives, which today represent a wealth of experiences and initiatives that we should be aware of, monitor and analyse.

Globally, there are many examples we can draw on. Platforms like the Open Government Partnership have seen success. The Dominican Republic is co-creating an AI Code of Ethics with government agencies, civil society organisations and technologists that will shape how AI is deployed and governed. Colombia is developing an AI governance model through a multi-stakeholder dialogue mechanism — civil society and academia is currently drafting a bill there. And Nigeria has committed to building an inclusive digital governance framework, with a multi-stakeholder coordination mechanism and a national policy on the ethical use of emerging technologies.

These initiatives, all rooted in the local context and all based on sharing power and decision-making with citizens, hold a promise: deliberation and citizen participation are possible. They do not hinder innovation; on the contrary, they foster the prospect of progress that is genuinely felt by the whole of society.

The next two years will be decisive. Across the globe, governments are making foundational decisions that will shape the deployment of AI for the coming decade. National strategies are moving from planning to implementation. Public procurementTransparency in the procurement process can help combat corruption and waste that plagues a significant portion of public procurement budgets globally. Technical specifications: Commitments that aim t... More frameworks are being established. Regulatory architectures are taking shape.

These decisions are being made now — with or without democratic safeguards. Once systems have been procured, suppliers locked in, and institutional structures established, correcting the course becomes exponentially more difficult. Investing early in transparent and participatory frameworks avoids costly corrections and builds public trust from the outset.

We have learnt this from social media: initial choices, if not regulated in time, create positions, power dynamics and economic models that are very difficult to change. The opaque systems deployed today become tomorrow’s infrastructure, and governance is reduced to damage control rather than design.

Human dignity at the centre

Ultimately, the issue goes beyond technical regulationGovernment reformers are developing regulations that enshrine values of transparency, participation, and accountability in government practices. Technical specifications: Act of creating or reforming .... It concerns what we wish to preserve: human dignity in the face of systems that could reduce us to profiles, scores and predictions. The value of work in the face of transformations that could render millions of people ‘useless’ in the eyes of efficiency algorithms. The very possibility of democratic deliberation in the face of tools capable of manipulating attention, fabricating parallel realities, and segmenting public debate into hermetic echo chambers.

The trajectory of AI must be shaped by global, national and local challenges, and by the interests of all — not just the private interests of big tech companies and their shareholders. Defining the ‘public interest’ of AI through open and inclusive conversations with civil society ensures that its development aligns with public values, builds trust, and prevents political capture and the concentration of power.

These risks are not distant dystopias. They are already here, in part, insidiously. And only open, inclusive, and rigorous governance can enable us to confront them.

In 2016, Paris asserted that 21st-century democracy would be open or it would not be at all. In 2026, we must go further: 21st-century democracy will be capable of governing artificial intelligence — or it will be governed by it.