As policy makers are grappling with developing appropriate regulation and measures to promote responsible AI, the Forum on Information and Democracy releases a paper exploring a novel idea to govern trustworthy AI: a voluntary certification mechanism for public interest AI. This idea was first presented in the Forum’s policy framework on AI as a Public Good: Ensuring Democratic Control of AI in the Information Space published in February 2024.
Incentive schemes are a powerful means to promote responsible and ethical behavior, as the legacy of the Fairtrade movement shows. Building upon this experience, the Forum on Information and Democracy explores the creation of a voluntary certification mechanism for public interest AI. Putting in place, this mechanism could encourage the development and deployment of systems in the public interest. It thus suggests a certification mechanism for AI systems directly accessible to users.
“The asymmetry of information between AI systems providers and the users of these systems can lead the general public to unchartered territories. With this voluntary certification mechanism, we aim to reduce this asymmetry by, eventually, providing users and citizens with clear signals on the trustworthiness and values of AI systems they are using,” explains Camille Grenier, Executive Director of the Forum on Information and Democracy
Elaborated with the support of the Future of Life Institute, the paper explores the opportunities of such a mechanism, its assessment criteria and governance structure. It is meant as a contribution to the debate on setting-up an AI certification mechanism – a mechanism that could contribute to ongoing initiatives such as UNESCO’s Ethical Impact Assessment and the OECD’s forthcoming Due Diligence Guidance for Responsible AI.
“Voluntary certification has shown success in commodity trade, it can also contribute to tackle the AI challenges in emphasizing responsible development and deployment and materialising commitments” highlights Marta Cantero Gamito, Professor of IT Law at the University of Tartu and Research Fellow at the Florence School of Transnational Governance at the EUI. Professor Cantero is at the origin of the idea and was one of the 14 members of an international working group that elaborated the policy framework AI as a Public Good: Ensuring Democratic Control of AI in the Information Space.
In the coming months, the Forum will promote the idea of a voluntary certification mechanism for public interest AI towards different stakeholders. We will notably encourage the States of the Partnership for Information and Democracy to explore this idea during the forthcoming ministerial meeting on Information, Democracy and AI to be held on 24 September 2024 in New York at the margins of the UN General Assembly. Taking place in February 2025, the AI Action Summit could be a prestigious venue to kick start the actual creation of this mechanism.
[/et_pb_text][/et_pb_column][/et_pb_row][et_pb_row _builder_version=”4.19.4″ _module_preset=”default” theme_builder_area=”post_content”][et_pb_column _builder_version=”4.19.4″ _module_preset=”default” type=”4_4″ theme_builder_area=”post_content”][et_pb_button button_text=”Read the Paper” _builder_version=”4.19.4″ _module_preset=”default” theme_builder_area=”post_content” button_alignment=”center” custom_button=”on” button_text_color=”#FFFFFF” global_colors_info=”{%22gcid-4120a231-ed7b-45ee-bac4-58fa79180f26%22:%91%22button_bg_color%22%93}” button_bg_color=”gcid-4120a231-ed7b-45ee-bac4-58fa79180f26″ button_bg_enable_color=”on” hover_enabled=”0″ sticky_enabled=”0″ min_height=”96px” button_url=”https://fid.azka-agency.com/wp-content/uploads/2024/09/FID-Public-Interest-AI-Sept-2024.pdf”][/et_pb_button][/et_pb_column][/et_pb_row][/et_pb_section]
One Response