“Today, citizens’ trust in public institutions is at an all-time low,” the report says. AI for the betterprepared by Ashoka Spain. Concretely, the text qualifies, in our country, Less than 40% of the population says they trust politics and the Administration public.
This also comes at a time when content generated by artificial intelligence is flooding social networks. Which, according to experts, threatens to destabilize democracies, both in the country and those around us.
In this context, the document published by Ashoka this week includes some key lessons from Tech for Humanity, the entity’s global initiative that “explores how technology can be designed and applied to benefit people and the planet“.
The document aims to “offer public administrations a strategic framework for adopting AI that strengthens democracy, improve the lives of citizens and strengthen collective intelligence. » And this is based on the direct experience of those who are already applying this technology “with ethical criteria and social impact”.
22 proposals
From Ashoka they claim to believe that humanity – and Spain in particular – faces a “historic opportunity” to bet on a technological vision that consists of increasing human capacity and not replacing it with artificial intelligence. And they emphasize: “That it does not concentrate power, but that it redistributes it; that she does not exclude, but that she listens more and better.”
This is why they created a sort of roadmap to “ensure that AI actually strengthens democracy and citizen participation”. Something which, they say, must be done from the very design of public policies.
Specifically, their proposal includes a total of 22 ideas or strategies that they believe, if followed, will build a healthier relationship between people and technology.
Democratic AI
Ashoka’s proposal is structured around several axes:
Democratic governance of AI
-
Create permanent advisory councils at every level of government, with representatives from civil society, academia, the private sector and social entrepreneurship, to ensure ongoing technical and social advice on AI.
-
Establish a central registry of algorithms for public use, mandatory for all administrations, with clear language and which includes the objective, the data used and the results of audits.
-
Create a ministerial-level advisory body in which civil society is formally represented and oversees the design and creation of the state register.
-
Include ethics committees and regular algorithmic audits in all AI policies, to detect biases, errors and possible inequality-generating effects.
-
Ensure that no decisions affecting people depend exclusively on AI.
Infrastructure and data serving the common good
-
Funding public AI and data infrastructure as a permanent public good that involves a long-term commitment, with independent governance, ensuring that the data and resulting models and algorithms are accessible, verifiable and reusable.
-
Guarantee universal and free access to public AI infrastructure and data for local governments, SMEs, social entrepreneurs, researchers and civil society organizations, with multilingual accessibility and disability adaptation.
-
Establish a national data standards framework for application based on FAIR principles.
-
Strengthen data communities and cooperatives made up of citizens, in which they play a leading role in their creation and governance.
-
Create a technological sovereignty fund which allows the financing of open source projects aimed at the common good and the resolution of problems with social impact.
Internal capacities and institutional learning
-
Create a guidance guide on the use of AI, ethics, democratic governance and algorithmic transparency in the public sector.
-
Introduce incentives and digital skills requirements into entrance tests, training programs and career plans for civil servants; as well as incentives to retain talent.
-
Establish permanent public-private social collaboration platforms for AI knowledge exchange, with the participation of social entrepreneurs and the AAPP.
Hiring and financing for impact
-
Contractually guarantee privacy, transparency and ethical use of data in public tenders and licensing agreements.
-
Fund AI civic innovation labs that enable citizens, experts and entrepreneurs to experiment with data and models in controlled environments and with public resources. Also facilitate the promotion of pilots developed with AI in the implementation of public policies.
-
Promote a diverse ecosystem of suppliers, preventing the economic offer from having greater weight in public calls for tenders and rewarding social innovation and the common good.
-
Generate a market of technology suppliers whose criteria favor social entrepreneurship.
Citizen participation enhanced by AI
-
Integrate AI into participatory processes to ensure fairness and make them more accessible. This includes: multilingual translation, easy reading, analysis of power imbalances, visualization of contributions from vulnerable communities and assisted moderation under human supervision, to amplify the voices of excluded and minority groups.
-
Implement AI to analyze and synthesize citizen contributions in mass consultation and deliberation processes, in order to extract useful information for decision-making.
-
Establish systematic monitoring mechanisms to identify social entrepreneurship projects with high impact potential and facilitate their integration into the design and development of public policies.
-
Enable new channels powered by generative AI to access information and disseminate the results of assemblies and other citizen processes in an accessible and engaging way.
-
Apply AI to the collection and analysis of social and economic data to support public policies based on verified evidence.
All this because, as the authors of the text remind us, “faced with an ongoing technological revolution, marked by the risk of increasing inequalities or losing democratic control, “AI also represents a great opportunity to rethink how we design public policies that are more inclusive, open and collaborative.”.