Can Technology Make Public Consultations More Inclusive?

Town halls, online surveys, citizen juries: public consultations have multiplied in recent years, yet many still miss the people they most affect, from shift workers and carers to rural residents and migrants. Governments are now betting on digital tools, AI translation, mobile-first platforms and data dashboards to widen participation and make feedback usable at scale. But can technology genuinely make consultations more inclusive, or does it risk hardening the digital divide, amplifying loud minorities and automating bias? The answer depends less on the gadgetry than on design, safeguards and trust.

Digital turnout rises, but so do gaps

Participation is the headline promise. When consultations move online, the cost of taking part drops, and evidence suggests turnout can climb sharply, particularly when processes are simple and mobile-friendly. In the United Kingdom, a parliamentary committee looking at “e-petitions” found that digital systems lowered barriers to entry and enabled rapid mobilisation, even if engagement still skewed toward already active citizens. Globally, the OECD has argued that digital participation can broaden reach, but only when it is paired with measures that address inequality in access and skills.

The gaps are stubborn. In 2024, the International Telecommunication Union estimated that roughly 2.6 billion people remained offline worldwide, and among those online, quality of access varies widely, from shared devices to unstable connections and expensive data. Even in high-income countries, digital exclusion tracks age, income, disability and geography, and consultations that assume constant connectivity can quietly filter out precisely the people policymakers say they want to hear from. Accessibility is another fault line, because if a platform is not compatible with screen readers, captioning or easy-language standards, the promise of “open to all” becomes little more than branding.

There is also the question of time, not just technology. Digital participation can compress a complex policy debate into a few clicks, and that format often favours those with spare time and confidence in bureaucratic language. If consultations are to be inclusive, platforms must let citizens engage in different ways, from quick polls to long-form submissions, and then make clear how those contributions will be read, coded and acted upon.

AI can translate voices, not legitimacy

Translation is where technology can make an immediate difference. Automatic speech-to-text, real-time captioning and machine translation can reduce linguistic barriers, and when they work well, they turn a monolingual process into something closer to a multilingual public square. For consultations involving minority languages, diaspora communities or cross-border issues, this matters, because language is often the first gatekeeper of civic life.

But AI cannot manufacture legitimacy. Machine translation still struggles with context, dialect, legal nuance and culturally specific references, and errors can distort meaning, particularly in testimonies about housing, policing or health where a single phrase can carry heavy implications. The risks increase when AI summaries replace human review, because summarisation models can compress varied perspectives into a falsely tidy consensus, or overrepresent comments that are longer, more emotional or easier for the model to parse. Researchers have repeatedly warned that algorithmic systems can encode bias present in training data, and that auditing and transparency are essential when automated tools affect civic outcomes.

In practice, the most credible approaches treat AI as assistive rather than authoritative. That means publishing plain-language explanations of how comments are processed, keeping human oversight in the loop for sampling and verification, and providing participants with a chance to correct transcripts or translations. It also means recognising that a consultation is not just data collection, it is a democratic ritual, and people want to feel listened to by humans, not merely ingested by software.

Identity checks fight fraud, and raise privacy stakes

Online consultations face a basic problem: how to prevent manipulation without scaring off legitimate participants. Bot-driven submissions, coordinated “astroturfing” campaigns and duplicate responses can overwhelm systems, and they are no longer hypothetical. Identity verification, device fingerprinting and anomaly detection can help maintain integrity, yet the more aggressive the checks, the higher the privacy cost, especially for groups with reasons to fear surveillance.

This is where the design choices become political. A platform that requires a government-issued ID may improve data quality, but it can exclude undocumented residents, people experiencing homelessness, or citizens who lack up-to-date documents. Conversely, a process that allows fully anonymous participation may welcome vulnerable voices, but it can also become easy to game, and public trust can erode if outcomes appear captured by organised groups. Some jurisdictions have tried hybrid models, allowing anonymous publishing but verified submission, or tiered participation where certain actions, such as voting on priorities, require stronger authentication than leaving an open comment.

Data minimisation is the inclusion-friendly principle too often ignored. Collect only what is necessary, store it briefly, and explain it clearly; otherwise participation becomes a trade, voice in exchange for exposure. This matters not only for privacy, but for safety, and it is especially relevant when consultations touch on sensitive issues such as domestic violence services, immigration enforcement or policing. People who most need a voice are often those most at risk if their identity leaks.

Inclusion needs offline bridges, not apps alone

Technology can widen the door, but it cannot replace the bridges that bring people to it. The strongest evidence from democratic innovation points toward “blended” approaches, combining digital tools with in-person outreach, facilitated workshops and targeted support. Citizens’ assemblies, deliberative mini-publics and community panels typically rely on recruitment methods that correct for demographic skews, often using stratified random sampling, childcare provision and compensation, because inclusion requires resources, not only interfaces.

Good platforms can, however, make those offline efforts more powerful. Mobile-first design helps when the smartphone is the primary device, and accessibility-by-default, including captions, adjustable text, and clear navigation, can reduce friction. Outreach can be improved through partnerships with libraries, community centres and trusted local organisations, which can host “assisted digital” sessions where staff help people submit feedback without taking over their voice. Even the smallest changes, such as keeping consultations open outside office hours, allowing voice notes, or offering call-in options, can shift who participates.

There is also a broader lesson from other areas of public life: people engage when the stakes are clear and the pathway is credible. If a consultation feels performative, participation drops, and the same communities that are underrepresented are often the first to disengage. That is why feedback loops are crucial, publish what was heard, show what changed, and explain what could not change, with reasons. For readers interested in how different jurisdictions structure participation and residency-linked civic incentives, the Vanuatu Golden Passport example illustrates how digital access, identity processes and administrative design can intersect, even if consultations themselves require separate safeguards.

Planning your next consultation, and its budget

Inclusivity is rarely free. Budget for translation, accessibility testing, moderation, outreach partners and independent evaluation, and set aside funds for participant compensation when deliberation demands time. Reserve venues and staff for offline sessions early, then align timelines so that in-person feedback enters the same pipeline as online responses, rather than being treated as an afterthought.

Where public funding is limited, seek grants for digital inclusion, disability access and civic innovation, and publish a short “participation plan” that explains choices on privacy, verification and AI use. Technology can help, but only when the process is designed to earn trust, protect the vulnerable and demonstrate impact.