The Paris Charter on Artificial Intelligence in the Public Interest

To achieve the potential benefits and preventing and mitigating the risks of emerging technologies for people and the planet, AI development, deployment and governance must be in the public interest. Public interest manifests differently for different communities, countries, and contexts, it requires opportunities for public participation, and it must serve equity and equality. We acknowledge that the mission and vision of artificial intelligence in the public interest builds on and is strengthened by existing definitions and academic research, public sector approaches, and civil society efforts. 

Neither the benefits nor the harms from technology fall on society in a proportionate manner. Benefits are typically available to the most advantaged in society and harms all too often are faced by the most disadvantaged. To that end we reaffirm our commitment to prevent and mitigate individual and collective harms, risks, threats and violations caused by the use and abuse of AI. We recognize that artificial intelligence should not be developed and deployed in areas where it is incompatible with international human rights law.  

In addition to appropriate safeguards to prevent, mitigate and remedy any adverse impact arising from the use of artificial intelligence, we need an affirmative vision to fully serve AI in the public interest. 

The benefits of AI in the public interest rely on building open public goods and infrastructure, providing an alternative to existing market concentration, ensuring democratic participation, enforcing accountability and developing environmentally sustainable solutions. To fulfil this vision, we focus on enabling conditions for infrastructure in areas with demonstrated benefits in the public interest.  

The barrier to scaling AI models has been assumed to be primarily the lack of availability and affordability of compute. While computational power is indeed subject to market concentration failures, the availability of high-quality data with adequate data governance structures proves to be the main bottleneck. New ways to access high quality data is key to unlock and ensure that public interest prevails. These new ways to access data must happen in compliance with the rights to privacy and data protection in line with national and international frameworks as essential safeguards and as the necessary precondition for people, companies and institutions to trust the AI ecosystem and enable necessary data flows. 

Smaller models can be designed to respond to concrete societal and community needs within specific local and cultural contexts, and given they require comparatively less computational power and data, will have a much less pernicious impact on the environment.      

With this in mind, we agree on the following principles: 

Openness drives progress in science, catalyzes innovation and enables competition. Today, openness in AI is largely driven by a few actors’ decision to partly open their foundation models. A resilient ecosystem is needed to support the development of open models, spanning both standard setting, tooling and best practices. 

Accountability across every step of AI design, development, and deployment is a cornerstone in achieving AI for the public interest. Accountability relies on the enforcement of existing national and international frameworks, enabling conditions for research, oversight and empowered institutions and civil society.

Participation and transparency are prerequisites for democratic governance in AI in the public interest.  

To achieve these objectives, we support the establishment of a global initiative for AI in the public interest. The undersigned country partners commit to uphold the values of AI in the public interest laid down in this Charter. We will address the challenges and realize the potentials for AI in the public interest through identifying our contributions and aligning on shared objectives in select areas of collaboration by 30 June 2025.

By enabling AI systems that are open, diverse, sustainable, locally relevant, development-centered and accessible around the world, and by providing structured access to high-quality datasets, models, compute and independent auditing and accountability tools, this initiative will serve as a platform for AI innovators, reformers, and advocates in their own countries while fostering international cooperation.

Through the initiative for AI in the public interest we commit to encourage a more comprehensive and inclusive design of AI in the public interest, in terms of technology, organization and institutions that serve different jurisdictions and communities in attaining similar success.

The initiative for AI in the public interest will enable civil society and impacted communities, academia, governments and other stakeholders to engage in a deliberation process to embed and enforce public interest within and beyond this initiative.

We will pursue this vision through collaboration, and it will enable participation to ensure that public interest prevails in the initiative’s mission, governance and impact.

Endorsed by: 
Chile 
Finland
France
Germany
India
Kenya
Morocco
Nigeria
Slovenia
Switzerland