A Founding Series No. 8 of 11; We've thought carefully about the ways this could fail. We think talking about them openly is part of how we prevent them.
Most companies, when they write about their work, do so from a posture of confidence bordering on inevitability. The product is revolutionary. The mission is clear. The team is exceptional. Risks are acknowledged briefly, then buried under a paragraph about the size of the market opportunity.
We are going to do the opposite.
This post is a catalog of the specific ways we believe SureVoter could go wrong — not as a theoretical exercise, but as a practical commitment to the kind of intellectual honesty that we think this mission requires. If we are asking the American public to trust us with civic infrastructure, then the least we can do is show them that we have stared clearly at our own failure modes.
There are seven we worry about most.
The tools SureVoter builds will involve algorithms — systems that take in information and produce structured outputs like ballot summaries, candidate comparisons, and constituent signal analyses. Algorithms, no matter how carefully designed, encode the assumptions of the people who build them.
A ballot summary tool trained primarily on data from urban media markets may produce subtly different quality outputs for rural ballot measures. A candidate profile system that relies on available digital records will disadvantage candidates who have less digital footprint — which often correlates with being less wealthy, less established, and serving smaller communities. A constituent signal tool that weights responses by engagement patterns may systematically under-represent older voters who engage with civic technology differently than younger ones.
None of these biases would be intentional. All of them could systematically distort the information environment in ways that advantage some voters and disadvantage others — and in ways that are difficult to detect without specifically looking for them. [1]
Our mitigation: We are committing to regular, independent algorithmic audits — conducted by researchers who are not on our payroll and who will publish their findings publicly. We are building diverse teams precisely because homogeneous teams build tools with homogeneous blind spots. And we are designing our platform to make its own methodology transparent and legible, so that users and external researchers can scrutinize how outputs are being generated.
This is not a complete solution. It is a posture of ongoing vigilance rather than confident arrival.
SureVoter is designed to amplify the signal of ordinary citizens over the noise of organized interest groups. But sophisticated actors who understand how our tools work will attempt to game them.
If our constituent signal tools respond to volume, well-resourced organizations will generate volume. If our candidate profile system relies on self-reported information, campaigns will learn to optimize their self-reporting. If our ballot summary methodology favors certain types of sources, interest groups will learn to produce content that those sources are likely to cite.
This is not paranoia. It is the predictable behavior of any organized interest that perceives SureVoter as a threat to its current informational advantages. The history of every information platform that has attempted to provide neutral civic signal — from fact-checkers to nonpartisan news aggregators — includes exactly this dynamic. [2]
Our mitigation: We are designing manipulation-resistance into the architecture from the beginning, rather than bolting it on after the fact. This includes weighted verification systems for constituent input, transparent sourcing requirements for candidate information, and ongoing red-team exercises in which we actively attempt to manipulate our own tools to find and close the vulnerabilities before bad actors do.
We will not win every round of this competition. We will lose some, learn, and adapt — publicly.
SureVoter will, if it works, reduce the informational advantages that certain political actors currently enjoy. Some of those actors are in positions of regulatory authority.
This is not hypothetical. Civic technology platforms that have disrupted existing political power arrangements have faced regulatory challenges, legislative threats, and in some international contexts, outright suppression. [3] In the United States, a platform that is perceived as threatening to the interests of a governing majority — at the federal or state level — could face hostile regulatory attention, selective enforcement of ambiguous statutes, or legislative attempts to impose requirements that compromise its independence.
Our mitigation: Our dual structure — private operating company plus independent nonprofit Foundation — provides some protection, as does our radical nonpartisanship. An organization that demonstrably serves voters of both parties equally is a harder target for partisan regulatory action than one that can be credibly characterized as ideologically aligned. We are also investing in legal resources specifically designed to defend the platform's independence, and building relationships with the broad coalition of civic organizations across the ideological spectrum that share an interest in citizen-aligned information infrastructure.
We are also working to ensure that our governance documents are legally robust enough to withstand challenges — not just morally clear enough to be defended in the court of public opinion.
Earlier in this series, we described how nonprofit organizations can drift toward the ideological preferences of their major donors over time — not through corruption, but through the entirely natural human dynamics of any organization that depends on the continued approval of funders.
The SureVoter Foundation is designed to steward the platform's civic integrity permanently. But the Foundation itself must be funded. And the sources of that funding, if not carefully designed, could introduce precisely the distortions the Foundation is meant to prevent. [4]
Our mitigation: We are designing the Foundation's endowment and funding model around diversification requirements — structural limits on the percentage of Foundation funding that can come from any single donor or category of donor. We are exploring structures that would give the Foundation a sustainable revenue base derived from the platform's commercial operations — reducing dependence on philanthropic funding over time without compromising the Foundation's independence from the operating company.
We do not yet have a complete answer to this. We are naming it as an open problem because it is one.
There is a version of SureVoter's success that we actively do not want: one in which the platform becomes so dominant in the civic information space that it creates a new single point of failure — or a new single point of control.
If tens of millions of Americans rely on SureVoter as their primary source of civic information, the platform's outputs become enormously consequential. A technical failure during a high-stakes election could deprive voters of information at a critical moment. A successful data breach could expose sensitive civic engagement data. A subtle methodological change — made in good faith but incorrectly — could influence election outcomes in ways that are never fully visible.
We are, in short, worried about becoming too important. [5]
Our mitigation: We are designing the platform's architecture with interoperability and openness as explicit goals — making it possible for other civic information tools to coexist with SureVoter rather than be displaced by it. We are committing to open-sourcing significant portions of our methodology so that independent researchers and alternative platforms can build on and verify our work. And we are building redundancy into our infrastructure specifically to prevent single points of failure during high-stakes civic moments.
A civic information ecosystem is healthier with multiple trusted actors than with a single dominant one — even if that single dominant actor has the best intentions.
Companies that begin with clear civic missions sometimes lose them — not through a dramatic betrayal, but through the accumulation of small compromises, each individually defensible, that together redirect the organization's center of gravity from its mission to its commercial survival.
The pressure points are predictable: a fundraising shortfall that makes a large donation from an ideologically interested donor hard to refuse; a product decision that would increase engagement but subtly reward partisan content; a partnership opportunity with a politically connected organization that offers distribution in exchange for influence; a talent decision that prioritizes growth experience over civic commitment.
None of these are hypothetical. They are the ordinary pressures of any organization operating in the real world. [6]
Our mitigation: The Foundation's structural authority over the operating company is specifically designed to create a binding constraint on mission drift — one that cannot be overridden by a board under commercial pressure or a CEO trying to meet quarterly targets. We are also committing to publishing an annual mission integrity report — an honest public accounting of the commercial pressures we faced in the previous year, how we responded to them, and whether we believe our responses were consistent with our founding mission.
This is accountability by design, not by aspiration.
The most humbling risk on this list is the simplest one: we might be building the wrong thing.
Our diagnosis of the problem — voter information gaps, legislative data fog, distorted constituent signal — is based on extensive research, expert input, and our own analysis of civic dysfunction. But we are not voters in every community we are trying to serve. We are not state legislators in under-resourced offices trying to navigate a hostile information environment. We are not the county clerk managing a local election with limited technology and staff.
The tools we design based on our current understanding of those experiences may not be the tools those people actually need. We may optimize for problems that are less important than we think, and miss problems that are more important than we realize. [7]
Our mitigation: We are building a sustained civic listening program — ongoing, structured engagement with the voters, officials, and election administrators whose experiences must shape our design decisions. We are treating this not as a marketing activity but as a core product development function. And we are committing to publishing what we learn — including the ways in which our original assumptions have been wrong.
The willingness to be corrected by evidence is, we believe, among the most important capabilities a civic technology organization can have.
A reasonable person might ask: why would a company publish a document cataloging its own potential failures?
The answer is that we don't think we have a choice — not if we want to be the kind of organization we are claiming to be.
The American public's trust in institutions is at a historic low. It did not get there because institutions failed dramatically and visibly. It got there through the accumulation of small deceptions — the reassurances given when reassurance wasn't warranted, the risks minimized when they should have been named, the problems discovered late because no one was looking for them.
We are asking voters and citizens to trust us with something important. The minimum we can offer in return is honesty about what could go wrong, and transparency about how seriously we are working to prevent it.
We don't expect to get everything right. We expect to learn, adapt, and remain accountable to the mission — in public, over time.
That is the commitment.
SureVoter builds with transparency because we believe trust is earned through honesty, not assurances. Follow our build at SureVoter.com.
Footnotes
[1] O'Neil, C. (2016). Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. Crown. O'Neil's analysis of algorithmic bias across civic and commercial domains provides the foundational framework for understanding how well-intentioned automated systems can produce systematically discriminatory outputs — particularly when training data reflects existing social inequalities.
[2] Wardle, C., & Derakhshan, H. (2017). Information Disorder: Toward an Interdisciplinary Framework for Research and Policy Making. Council of Europe. Documents the systematic strategies through which organized actors exploit information platforms — including fact-checking services, aggregators, and civic information tools — through coordinated inauthentic behavior designed to produce artificially amplified signals.
[3] Freedom House. (2023). Freedom on the Net: The Repressive Power of Artificial Intelligence. See also: Deibert, R. (2020). Reset: Reclaiming the Internet for Civil Society. House of Anansi Press. While the most acute examples of civic technology suppression occur in authoritarian contexts, the report documents pressure on civic information platforms in democratic systems including selective regulatory enforcement and legislative threats.
[4] Callahan, D. (2017). The Givers: Wealth, Power, and Philanthropy in a New Gilded Age. Knopf. The documented pattern of philanthropic influence on nonprofit mission is extensive and well-established. Callahan's analysis finds that even foundations with explicit commitments to nonpartisanship have drifted toward the ideological preferences of major donors over multi-decade periods.
[5] Zittrain, J. (2008). The Future of the Internet — And How to Stop It. Yale University Press. Zittrain's framework for thinking about centralization risks in information infrastructure — the danger that dominant platforms create dependency that ultimately undermines the resilience and diversity of the systems they replace — is directly applicable to civic information platforms.
[6] Stacy Mitchell & Olivia LaVecchia. (2016). Amazon's Stranglehold: How the Company's Tightening Grip Is Stifling Competition, Eroding Jobs, and Threatening Communities. Institute for Local Self-Reliance. Cited here not for its specific Amazon analysis but for its framework of how mission drift occurs through commercial scaling — a dynamic that applies across sectors including civic technology.
[7] Costanza-Chock, S. (2020). Design Justice: Community-Led Practices to Build the Worlds We Need. MIT Press. Costanza-Chock's framework for participatory design — the principle that the people most affected by a technology must be central to its design, not merely consulted after the fact — directly informs our approach to civic technology development.