Over time, more and more parliaments have embraced digital tools and services. The rise of AI is expected to further accelerate this trend and play a significant role in transforming legislatures from paper-driven organisations into data-driven institutions. These guidelines aim to prepare representative institutions for the introduction and use of AI in the parliamentary workspace. This publication was drafted by an international group of parliamentary scholars and professionals over a period of 8 months, from September 2023 to April 2024, building upon earlier work in the field. We recognise that the guidelines remain a work in progress within an ever-changing technological and institutional context.
That said, the guidelines have the potential to ensure informed regulation to empower parliaments across policy formulation, public engagement, capacity building, and more. They can help ensure the responsible integration of AI, addressing transparency and ethics in political and administrative processes, thus strengthening public trust, and safeguarding the public interest. Additionally, these guidelines can help align AI tools and services with democratic principles and societal needs. They also contribute significantly to the sharing of best practices and ethical conduct, ultimately supporting knowledge growth and collaboration among the parliamentary community. These guidelines are relevant for local, regional, national, and supranational parliaments in a multilevel governance context. They take a holistic approach, addressing ethics, privacy, security, oversight, system design, and education.
They look into specific aspects of the use of AI in parliaments, including scope, examples, and factors that are critical for successful implementation. This makes them both useful for tackling contemporary issues and relevant for assessing more theoretical ones, such as the implications of artificial general intelligence (AGI) for legislatures. Technology moves rapidly. Hence, the guidelines were designed to be technologically agnostic – in other words, they do not address any specific AI technology. However, indications of major technology trends are outlined, such as generative and hybrid AI. We hope this publication is disseminated as widely as possible, to reach every parliament, parliamentarian, administrator, and anyone else genuinely interested in maximising the positive effects of AI in legislatures while minimising the potential risks.
It is for exactly this reason that the editors and the other authors are committed to cooperating further with parliamentary and societal stakeholders to drive the further development of these guidelines. Communication of the guidelines, collaboration, and customisation. will help ensure effective implementation and adaptation in diverse institutional contexts. We welcome proposals from those interested in working with us to translate the guidelines, develop training materials, provide support for their implementation, or share best-practices to accelerate their effective integration into the parliamentary workspace. Proof-of-concept and pilot projects, whether unilateral, bilateral, or multilateral, will allow for practical testing and refinement of the guidelines in diverse contexts and we look forward to learning from them.
Fotios Fitsilis, Hellenic Parliament, Greece
Jörn von Lucke, Zeppelin University, Germany
Franklin De Vrieze, Westminster Foundation for Democracy, United Kingdom