Around the world, people are confronting a new kind of uncertainty. Parents trying to help children navigate online harms. Workers worried about job losses from automation. Women in politics facing artificial intelligence (AI)-driven abuse. Citizens trying to understand which information they can trust. In these moments, AI is not an abstract technological shift; it is a governance challenge. 

Parliaments sit at the centre of this challenge. As legislators, overseers and representatives, they are uniquely placed to make sure AI serves the public good by setting guardrails, scrutinizing executive action and protecting those most at risk. This comes at a moment when UNDP’s Human Development Report (HDR) warns that AI’s impact on human development will depend less on what the technology can do and more on the political choices we make now. For parliaments, the passive era of 'wait and see' is over. The task now is to move to anticipatory governance.

UNDP supports one in three parliaments worldwide, including in helping them navigate digital transformation and govern emerging technologies in ways that are inclusive, rights-based and accountable.  

A question of cooperation

Parliaments face a similar dilemma: how to ensure that technological breakthroughs do not concentrate power in the hands of a few? 

This is precisely why the Global Digital Compact calls for a shared vision on digital governance grounded in equity, human rights and inclusive participation. AI governance cannot be shaped by technical experts or private actors alone; it must reflect democratic debate, public accountability and meaningful engagement with citizens. Parliaments are a crucial space where these conversations converge. 

UNDP’s support reflects this reality. From digital readiness assessments in Thailand, AI trainings in Armenia and Uruguay, to peer-learning platforms such as the UNDP-IPU Expert Group on Parliament Engagement on Digital Policy, we help parliaments strengthen their core functions of lawmaking, representation, scrutiny and oversight so that cooperation becomes an institutional habit.

The power to ask hard questions

One lesson emerging from UNDP’s engagement with parliaments in over 60 countries is the lack of AI literacy and confidence to interrogate AI systems and their impacts. 

AI literacy for parliamentarians is not about learning to code but about learning to question. Parliamentarians benefit most when they ask the right questions about, among other things, data sovereignty, vendor lock-in, cloud jurisdiction and by understanding how algorithms affect everything from teenagers’ attention spans to the spread of disinformation. This kind of literacy is not a one-off exercise. It is like taking the institutional brain to the gym; it requires routine practice, continuous learning and constant questioning as technologies evolve. 

From replacement to complementarity

For too long, the public narrative on AI has been dominated by fears of replacement. Parliaments are beginning to reframe this debate by looking at where AI can sensibly support existing work, and where human judgement must remain central. Already, many parliaments are experimenting with pilots such as AI chatbots, transcription tools and research assistants.

However, without institutional architecture, pilots remain islands of innovation that do not change how a parliament works. Pilots alone do not transform institutions. Without budget lines, defined roles, ethical guidelines and human oversight, tools remain fragmented and unevenly adopted.

Mauritius offers a clear success story. With UNDP’s support, the National Assembly now uses AI to transcribe debates, summarize long sessions and make information available in multiple languages. These tools help members of parliament and citizens access information faster, while centring human judgement and oversight.

Start where people are hurting most

When parliaments engage with AI, the most meaningful conversations should not start with technical systems but with people. They need to begin in the places where citizens are already feeling the pressure of digital transformation; farmers facing climate shocks, families struggling to access public services or communities exposed to online harm.

Instead of asking “What can AI do?”, parliaments need to ask, “Whose problems should AI help solve first and under what conditions?” This lens shifts attention toward equity, protection and dignity, ensuring that innovation does not race ahead of the realities lived by those most at risk of being left behind. 

AI-driven harassment and tech-facilitated gender-based violence (TFGBV) are increasingly undermining participation in public life, particularly for women in politics, activists, journalists and minority groups. These are lived harms that silence voices, distort representation and weaken democracy. Through UNDP’s eMonitor+ tool, now deployed in dozens of countries, parliaments and electoral actors are analyzing online hate speech and TFGBV in real time to inform prevention strategies. 

Choosing the future together

The HDR reminds us that the future of AI does not have a pre-written script; it is “a matter of choice”. It’s important for parliaments to recognize that they cannot approach AI as a purely technical issue. It is about people, trust in institutions, protection of rights and the kind of societies we want future generations to inherit.

This commitment to shared responsibility has begun to take shape through emerging spaces of parliamentary collaboration. A recent conference co-organized by UNDP, the Commonwealth Parliamentary Association and the Inter-Parliamentary Union brought together over 200 parliamentarians and experts from around 70 countries. From this process emerged the Kuala Lumpur Declaration on Parliaments and Responsible AI. Through it, parliaments are choosing to build literacy, redesign institutions, strengthen oversight and engage globally on norms and standards. 

UNDP’s role is to stay alongside parliaments in that journey as a convener, a technical partner and a champion of people-centred and rights-based AI. The window to shape AI in the public interest is still open, but it will not remain so for long. By moving from reactive fear to proactive, human-centred governance, parliaments can ensure that AI strengthens democracy rather than weakening it.

If parliaments can keep that spirit of urgency and cooperation alive, AI can become not another fault line in our democracies, but a tool to renew them.