Category Archives for public services

Agents for the few, queues for the many – or agents for all? Closing the public services divide by regulating for AI’s opportunities.

(co-authored with Luke Jordan, originally posted on Reboot Democracy Blog)

Inequality in accessing public services is prevalent worldwide. In the UK, “priority fees” for services like passport issuance or Schengen visas allow the affluent to expedite the process. In Brazil, the middle-class hires “despachantes” – intermediaries who navigate bureaucratic hurdles on their behalf. Add technology to the mix, and you get businesses like South Africa’s WeQ4U, which help the privileged sidestep the vehicle licensing queues that others endure daily. An African exception? Hardly. In the U.S., landlords use paid online services to expedite rental property licensing, while travelers pay annual fees for faster airport security screening.

If AI development continues and public sector services fail to evolve, inequalities in access will only grow.  AI agents – capable of handling tasks like forms filling and queries – have the potential to transform access to public services. But rather than embracing this potential, the public sector risks turning a blind eye – or worse, banning these tools outright – leaving those without resources even further behind.

The result? The private sector will have to navigate the gaps, finding ways to make AI agents work with rigid public systems. Often, this will mean operating in a legal grey zone, where the agents neither confirm nor deny they are software, masquerading as applicants themselves. Accountants routinely log into government tax portals using their clients’ credentials, acting as digital proxies without any formal delegation system. If human intermediaries are already “impersonating” their clients in government systems, it’s easy to envision AI agents seamlessly stepping into this role, automatically handling documentation and responses while operating under the same informal arrangements.

The high costs of developing reliable AI agents and the legal risks of operating in regulatory grey zones will require them to earn high returns, keep these tools firmly in the hands of the wealthier – replicating the same inequalities that define access to today’s analogue services. 

For those who can afford AI agents, life will become far more convenient. Their agents will handle everything from tax filings to medical appointments and permit applications. Meanwhile, the majority will remain stuck in endless queues, their time undervalued and wasted by outdated bureaucratic processes. Both groups, however, will lose faith in the public sector: the affluent will see it as archaic, while the underserved will face worsening service as the system fails to adapt.

The question is no longer whether AI agents will transform public services. They will. The partners of Y Combinator recently advised startup founders to “find the most boring, repetitive administrative work you can and automate it”. There is little work more boring and repetitive than public service management. The real question is whether this transformation will widen the existing divide or help bridge it. 

Banning AI agents outright is a mistake. Such an approach would amount to an admission of defeat, and entrenching inequalities by design. Instead, policymakers must take bold steps to ensure equitable access to AI agents in public services. Three measures could lay the groundwork:

  1. Establish an “AI Opportunities Agency”: This agency would focus on equitable uses of AI agents to alleviate bureaucratic burdens. Its mandate would be to harness AI’s potential to improve services while reducing inequality, rather than exacerbating it. This would be the analogue of the “AI Safety Agency”, itself also a necessary body. 
  2. Develop an “Agent Power of Attorney” framework: This framework would allow users to explicitly agree that agents on an approved list could sign digitally for them for a specified list of services. Such a digital power of attorney could improve on existing forms of legal representation by being more widely accessible, and having clearer and simpler means of delegating for specific scopes.
  3. Create a competitive ecosystem for AI agents: Governments could enable an open competition in which the state provides an option but holds no monopoly. Companies that provided agents which qualified for an approved list could be compensated by a publicly paid fixed fee tied to successful completions of service applications. That would create strong incentives for companies to compete to deliver higher and higher success rates for a wider and wider audience.

A public option for such agents should also be available from the beginning. If not, capture will likely result and be very difficult to reverse later. For example, the IRS’s Direct File, launched in 2024 to provide free tax filing for lower-income taxpayers, only emerged after years of resistance from tax preparation firms that had long blocked such efforts – and it continues to face strong pushback from these same firms.

One significant risk with our approach is that the approval process for AI agents could become outdated and inefficient, resulting in a roster of poorly functioning tools – a common fate in government, where approval processes often turn into bureaucratic roadblocks that stifle innovation rather than enable it.

In such a scenario, the affluent would inevitably turn to off-list agents provided by more agile startups, while ordinary citizens would view the initiative as yet another example of government mismanaging new technology. Conversely, an overly open approval process could allow bad actors to infiltrate the system, compromising digital signatures and eroding public trust in the framework.

These risks are real, but the status quo does nothing to address them. If anything, it leaves the door wide open for unregulated, exploitative actors to flood the market with potentially harmful solutions. Bad actors are already on the horizon, and their services will emerge whether governments act or not.

However, we are not starting from scratch when it comes to regulating such systems. The experience of open banking provides valuable lessons. In many countries, it is now standard practice for a curated list of authorized companies to request and receive permission to manage users’ financial accounts. This model of governance, which balances security and innovation, could serve as a blueprint for managing digital agents in public services. After all, granting permission for an agent to apply for a driver’s license or file a tax return involves similar risks to those we’ve already learned to manage in the financial sector.

The path ahead requires careful balance. We must embrace the efficiency gains of AI agents while ensuring these gains are democratically distributed. This means moving beyond the simple dichotomy of adoption versus rejection, toward a nuanced approach that considers how these tools can serve all citizens.

The alternative – a world of agents for the few, and queues for the many – would represent not just a failure of policy, but a betrayal of the fundamental promise of public services in a democratic society.

Civic Tech and Government Responsiveness

For those interested in tech-based citizen reporting tools (such as FixMyStreet, SeeClickFix), here’s a recent interview of mine with Jeffrey Peel (Citizen 2015) in which I discuss some of our recent research in the area.

 

What shapes citizens’ evaluations of their public officials’ accountability? Evidence from local Ethiopia

 

I just came across an interesting paper by Sebastian Jilke published in Public Administration and Development. on the effects of access to information and participatory planning on citizens’ perception of local public officials. Below the summary of the paper:

In this article, we study which institutional factors shape citizens’ views of the local accountability of their public officials. Our departing assumption is that evaluations of local accountability do not merely reflect citizens’ political attitudes and beliefs, but also whether local institutions contribute to an environment of mutual trust, accountability and ultimately democratic legitimacy. Combining public opinion data from a large-N citizen survey (N=10,651) with contextual information for 63 local governments in Ethiopia, we look at access to information, participatory planning and the publicness of basic services as potential predictors of citizens’ evaluations of local public officials. Our findings suggest that local context matters. Jurisdictions that provide access to information on political decision-making are perceived to have more accountable officials. Moreover, when local governments provide public fora that facilitate citizens’ stakes in local planning processes, it positively affects citizens’ evaluations of the accountability of their officials. Our study adds to the  empirical literature by showing that establishing local institutions that can foster citizen-government relations at the local level through inclusive processes is crucial for improving public perceptions of accountability.

And a few more excerpts from the conclusion:

We have presented an empirical test of local institutional factors – particularly access to information,  participatory planning and publicness of basic services – and their impact on citizens’ perceptions of local accountability in Ethiopian local governments. Our empirical results show that two out of the three factors matter. Once a jurisdiction adopts participatory planning and/or provides access to information on political decision-making, it positively affects the way in which citizens perceive the accountability of their officials. In sum, both factors are thought to improve the relationship between citizens and their respective local governments. Hence, our findings suggest that establishing local institutions that can foster citizen-government relations at the local level are crucial for improving public attitudes towards local government. Furthermore, positive attitudes towards local government, furthermore, strengthen the democratic legitimacy of the state at the local level. Thus development practitioners and policy-makers may take these institutional factors into account when reforming local governments.

You can read an ungated version of the paper here [PDF].

And you can read more about the benefits of citizen participation here. 

Citizen Engagement Improves Access to Public Goods in Mexico

A paper recently published in World Development brings new and fascinating evidence from Mexico of the impact of participatory governance mechanisms on access to services.

Below are a few excerpts from the paper by Diaz-Cayeros, Malagoni, and Ruiz-Euler “Traditional Governance, Citizen Engagement, and Local Public Goods: Evidence from Mexico” (emphasis are mine):

The goal of this paper is to assess the effects of traditional governance on local public good provision. We ask whether poor indigenous communities are better off by choosing to govern themselves through “traditional” customary law and participatory democracy, versus delegating decisions concerning the provision of public goods to “modern” forms of representative government, structured through political parties. This is a crucial question for developing countries seeking to enhance accountability, and a central problem in the theory of participatory democracy.

Our research design takes advantage of an important institutional innovation in the state of Oaxaca, Mexico, that in 1995 allowed indigenous communities to choose their forms of governance. The reform gave full legal standing to a form of traditional indigenous governance called usos y costumbres (usos hereafter), which entails electing individuals to leadership positions through customary law in non-partisan elections, making decisions through participatory democracy, and monitoring compliance through a parallel (and often informal) system of law enforcement and community justice. If they did not choose usos, municipalities could opt instead for party governance, which entails the selection of municipal authorities through electoral competition among political parties and the adjudication of conflicts only through the formal institutional channels, namely the state and federal judiciary.

(…)

Our results show that electricity provision increased faster in those municipalities governed by usos. They also suggest that traditional governance may improve the provision of education and sewerage. With respect to citizen engagement and elite capture, contrary to existing scholarly work, we find no evidence of entrenchment of local bosses (caciques) associated with the former ruling party, the Institutional Revolutionary party (PRI) in places ruled by usos. Our findings suggest that traditional participatory forms of governance do not handicap democratic development. Furthermore, municipalities governed by usos are more likely to hold open council meetings allowing citizens to participate in decisionmaking processes. We attribute better public goods coverage to differences in local governance and collective decisionmaking practices. We suggest three specific channels through which traditional governance affects local public good provision: the social embeddedness of municipal presidents, broader civic engagement in collective-decision making, and credible social sanctions. We argue that traditional governance practices (which include in our setting decision-making through direct participatory practices, the obligation to provide services for the community, and the establishment of a parallel system of justice), allow poor communities to better hold their political leaders accountable, prevent elite capture, and monitor and sanction non-cooperative behavior.

(…)

Systems of governance based on electoral competition among political parties differ essentially from usos because decisions are taken by politicians without an ongoing process of consultation with the citizenry. The monitoring and sanctioning dynamics that come into play when citizens gather in public assemblies are usually absent in party-run municipalities, and thus the allocation of resources for public goods seems sub-optimal.

(…)

Differences between the two types of governance that we presented in the paper point to a broader discussion of the organization of democracy. The delegated format of decision-making in electoral democracies dominated by political parties seems to bear a higher risk of agency loss than deliberative decision-making of what is often referred to as participatory democracy. (…) there are lessons to be extracted from the fact that, with regard to the provision of some basic services, a non-partisan political arrangement presented some advantages over the widespread electoral and party-based democratic organization. Participation and collective monitoring of authority are hugely important to maximize collective well-being.

Read the full paper here [PDF].

 

High-Frequency Data Using Mobile Phones: Incentives and Accountability

Great research note  [PDF] by Croke et al. (2012). Here’s the abstract:

As mobile phone ownership rates have risen dramatically in Africa, there has been increased interest in using mobile telephones as a data collection platform. This note draws on two largely successful pilot projects in Tanzania and South Sudan that used mobile phones for high-frequency data collection. Data were collected on a wide range of topics and in a manner that was cost-effective, flexible, and rapid. Once households were included in the survey, they tended to stick with it: respondent fatigue has not been a major issue. While attrition and nonresponse have been challenges in the Tanzania survey, these were due to design flaws in that particular survey, challenges that can be avoided in future similar projects. Ensuring use of the data to demand better service delivery and policy decisions turned out to be as challenging as collecting the high-quality data. Experiences in Tanzania suggest that good data can be translated into public accountability, but also demonstrate that just putting data out in the public domain is not enough. This note discusses lessons learned and offers suggestions for future applications of mobile phone surveys in developing countries, such as those planned for the World Bank’s “Listening to Africa” initiative.

Of particular interest to me is the fact that part of the design used financial incentives as a means to reduce nonresponse and attrition rates. In the technology and development world there has been lots of talk about “incentives to participate”, where the practical shortcut is often the provision of financial incentives. In Tanzania, for instance, the authors report that “respondents who successfully completed an interview were rewarded with an amount varying from $2 to $4”, not a negligible sum in the Tanzanian context.

Interestingly, in the working paper [PDF] from which this note is drawn, a footnote sheds some light on how effective these rewards were:

Remarkably in both Sudan and Tanzania the amount of the reward did not have a discernable impact on response rates.

But these findings are not as surprising as they may seem. Indeed, there is a good deal of evidence from behavioural economics pointing out that financial incentives might not work as well as traditional economics (and economists) would predict.

And a noteworthy excerpt on the limits of transparency and the role of existing institutions and actors:

One lesson is that  providing citizens with relevant, timely, and accurate data  about the actions of politicians, policy makers, and public service providers is not sufficient. For the data to have impact, they need to be accessible and disseminated widely, and in ways that allow them to be utilized by already existing institutions and actors.

This is an interesting point, although I am not sure to what extent existing institutions are enough. In the field of technology and governance, I believe that it has become quite clear that very little is achieved when technological solutions are not coupled with institutional innovations.

But that’s another story. In any case, a great read, and the type of effort that is badly needed in this space.

Local Environment and Monitoring of Public Health Service Delivery

picture by Dave Proffer on flickr

 

When is Community-Based Monitoring Effective? Evidence from a Randomized Experiment in Primary Health in Uganda 

By Martina Bjorkman and Jakob Svensson (2012)

Excerpts:

Access to quality services has been recognized as fundamental for wellbeing and economic development. However, in Africa and other developing countries, service delivery is often poor or nonexistent. Many argue that government bureaucracies may be ill equipped and lack incentives to improve the quality of public services. In response, development practitioners have started to experiment with involving beneficiaries in monitoring public service delivery and making service providers accountable to users. How best to design such interventions, and the impact of them, have been addressed in a handful of recent randomized field experiments. The results, to date, are mixed. While Banerjee et al. (2008) and Olken (2007) report minor or no effects on learning outcomes (in a project in primary education in India) and on corruption (in a road construction project in Indonesia), Bjorkman and Svensson (2009) and Duflo et al. (2009) report large positive improvements on average in a primary health intervention in Uganda and a primary schooling intervention in Kenya, respectively. What can explain these diverging findings? And more specifically, to what extent does the local sociopolitical environment influence users ability and willingness to monitor public service providers?

Using data from Bjorkman and Svensson (2009), linked to recently assembled data on ethnic and linguistic composition at the sub-national level for Uganda (Alesina and Zhuravskaya, 2008), and income data from the Uganda National Household Survey 2005 (UNHS, 2005), we test whether social heterogeneity, in income and ethnicity, can explain why some communities managed to push for better health service delivery while others were less successful. The results suggest that income inequality and, particularly, ethnic fractionalization adversely impact collective action for improved service provision.

***

As policymakers in developing countries search for ways to improve health and education for the poor, it is becoming clear that more is required than just additional funds. A key obstacle to better public services looks to be the weak incentives that providers face. Schools and health clinics are not open when they should be. Teachers and health workers are frequently absent from schools and clinics, and even when there, they spend significant time not serving the intended beneficiaries. Equipment, even when working, is not used. Drugs are misused, and public funds are expropriated. In response, a growing number of experts argue that more emphasis must be placed on strengthening beneficiary control that is, strengthening providers’ accountability to citizens/clients.

While there is evidence that such an approach can have large positive effects on service provision, there is also evidence of signiÖcant variation in outcomes. Using data from a randomized experiment in Uganda, we show that social heterogeneity, and specifically ethnic fractionalization, adversely impact collective action for improved service provision. As a result, the intervention resulted in a smaller increase in the quantity of primary health care provision in heterogeneous communities.

Our results have implications for both the design and evaluation of interventions aimed at strengthening beneficiary control in public service delivery programs. On program design, interventions should be adjusted to the local sociopolitical situation. As little is known about how this is to be done, our results open up an important agenda for research: How to enhance collective action in socially heterogeneous communities. On evaluation, ideally the researchers should design the evaluation protocol so as to be able to assess the impact conditional on the sociopolitical environments.

Read full study here [PDF].