The Academic Edition
What the latest research tells us about AI in parliaments
BY BEATRIZ REY
In the past two years, parliaments have moved from asking whether AI should play a role in lawmaking to testing how it already does. A new wave of academic research is tracing this transformation, revealing that the benefits and pitfalls of AI in legislative politics lies less in automation itself than in what it means for trust, legitimacy, and representation.
In the UK and Japan, Steven Pickering and colleagues examine how citizens view the use of artificial intelligence in parliamentary decision-making. Based on original surveys, they find that people are generally cautious: they are comfortable with AI assisting Members of Parliament but strongly oppose letting AI or robots make decisions instead of them (especially in the UK). Trust emerges as a key factor: confidence in government (and broader social trust in Japan) increases acceptance, while fear of AI reduces it. Demographic differences also matter – women and older respondents are more skeptical – and ideological effects vary by country.
In one experiment, Sarah Kreps and Maurice Jakesch investigate how AI language models such as GPT-3 could reshape communication between legislators and citizens. Facing high volumes of constituent messages, legislators often rely on generic responses; AI tools could make replies more efficient and personalized. Through two experiments comparing human-written, AI-assisted, and fully AI-generated correspondence, the authors find that AI with human oversight can actually increase constituent trust and satisfaction, outperforming impersonal auto-responses. However, poorly designed or undisclosed AI use may undermine confidence in legislators. The study concludes that the success of AI-mediated communication in politics depends on transparency, disclosure, and maintaining human accountability throughout the process.
Meanwhile, Viktor Suter and colleagues trace how AI is framed in parliamentary debates across four legislatures (the US Congress, European Parliament, Parliament of Singapore, and Swiss Federal Assembly) between 2014 and 2024. Using natural language processing and qualitative analysis, the authors find a sharp rise in AI-related discussions after the public release of ChatGPT in 2022, with debates dominated by themes of ethics, regulation, and governance. National priorities, however, diverge: the US frames AI around defense and security, Singapore around innovation and labor, the EU around ethical oversight, and Switzerland around regulatory caution. The study shows how geopolitical context and policy traditions shape how parliaments interpret and respond to emerging technologies.
In Italy, Ylenia Maria Citino documents how the Italian Chamber of Deputies has become one of the first parliaments to integrate generative AI (GenAI) into its legislative procedures. It analyzes how AI tools are used to process amendments and assist parliamentary staff, highlighting both the efficiency gains and the constitutional risks such technologies pose. The author argues that GenAI challenges core democratic principles — neutrality, transparency, and legitimacy — and also transforms the less visible, informal dimensions of parliamentary work that sustain deliberation.
Across contexts, AI is forcing legislatures to ask what parts of representation can be automated and what must remain human. As parliaments test algorithms in drafting, deliberation, and communication, scholars are mapping the same frontier politicians are now crossing in real time.
AI News
The Economic Community of West African States (ECOWAS) Parliament recently called for leveraging AI to enhance legislative drafting, citizen engagement, and democratic governance across West Africa. Lawmakers and regional leaders emphasized that artificial intelligence can strengthen efficiency and accountability if harnessed ethically and inclusively.
The US House of Representatives will launch a pilot program giving 6,000 staffers access to Microsoft Copilot, reversing a previous ban and signaling a new phase in Congressional modernization. Speaker Mike Johnson [R, LA] said the AI tool will help streamline work and improve constituent services.
The UK Parliament has released new guidance on the safe use of artificial intelligence for Members and staff of both Houses. Approved by the Commons and Lords committees, the document offers practical advice and will be regularly updated as AI technologies evolve.
As per the Inter-Parliamentary Union (IPU) Innovation Tracker, the House of Commons of Canada is advancing a structured, ethical AI strategy grounded in human-centered values such as accountability, inclusivity, and transparency. Since 2024, it has created dedicated AI committees, launched staff training and awareness campaigns, and tracked sentiment showing rising support for AI adoption (from 60% to 74% positive).
Highlights from the Sixth World Conference of Speakers of Parliament
The official publication from the Sixth World Conference of Speakers of Parliament is out now. Held at the Palais des Nations in Geneva (29–31 July 2025), with the 15th Summit of Women Speakers as its prelude, the gathering brought together parliamentary leaders from across the globe to reimagine multilateralism, accountability, and democracy for a changing world. Discover how parliaments are shaping peace through parliamentary diplomacy, amplifying citizens’ voices within the United Nations system, and turning global commitments into local action.
Events
November 17-19: Information Seminar on the Structure and Functioning of the IPU for Parliaments in Latin America and the Caribbean (virtual)
November 28-30: The Role of Parliament in Shaping the Future of Responsible AI, Inter-Parliamentary Union (Kuala Lumpur, Malaysia)
ModParl Reader Snapshot
We’d love to know more about our readers! This quick two-question survey helps us understand who’s reading ModParl and where you’re based.
Modern Parliament (“ModParl”) is a newsletter from POPVOX Foundation that provides insights into the evolution of legislative institutions worldwide. Learn more and subscribe at modparl.substack.com.
