The relationship between the American public and the technology sector has long been framed as a dynamic of external pressure: Congress summons CEOs for hearings, regulators draft rules, and advocacy groups campaign for change. But what happens when the machinery of government itself is run—and increasingly, owned—by the very companies it is supposed to oversee? In a revealing conversation on the Geek in Review podcast, legal scholar Hannah Bloch-Wehba pulls back the curtain on a system where Big Tech’s grip on government is not merely influential but structurally embedded. Her research suggests that the line between public authority and private enterprise has become so blurred that tech companies are no longer just vendors to the state; they are, in many cases, the state’s operational backbone. This article examines the hidden mechanisms Bloch-Wehba identifies, from AI regulation loopholes to police surveillance infrastructure, and explores what it means for public accountability when Silicon Valley calls the shots from inside the machine.
Beyond the Vendor Contract: How Tech Became the Shadow State
The prevailing assumption about government procurement is that agencies buy products or services from private companies, and those companies remain external actors subject to oversight. Bloch-Wehba’s work dismantles this notion. She argues that the relationship has evolved into something far more intimate and problematic. Tech companies do not simply park outside the gates of government, waiting for instructions. They are integrated into the daily operations of public agencies, from the Department of Homeland Security to local police precincts. This integration is not incidental; it is a deliberate design that allows private firms to exercise public power without the constitutional and statutory constraints that bind government actors.
One of the most striking points Bloch-Wehba makes is that this arrangement often bypasses standard accountability mechanisms. When a government agency uses a privately developed algorithm to determine bail eligibility, who is responsible if the algorithm produces biased outcomes? The agency can claim it relied on a technical tool. The company can claim it merely provided a neutral product. In this gap, accountability evaporates. Bloch-Wehba refers to this as a form of “foam” between public and private spheres—a permeable barrier that allows liability to slip away while the state expands its surveillance and enforcement capabilities.
The pandemic accelerated this trend. As government agencies rapidly digitized services and law enforcement adopted cloud-based analytic tools, the dependency on companies like Amazon, Microsoft, and Palantir deepened. Bloch-Wehba notes that these firms have become so entrenched that they are effectively designing the architecture of modern governance. This is not hyperbole; it is the logical endpoint of years of outsourcing, data sharing, and the adoption of proprietary algorithms for core public functions.
What Are the Constitutional Limits on AI in Government? A Direct Answer
The constitutional limits on AI use by government agencies are currently undefined and largely untested. Courts have applied existing doctrines—such as due process, equal protection, and the Fourth Amendment’s protection against unreasonable searches—to cases involving algorithmic decision-making, but there is no comprehensive legal framework that specifically addresses artificial intelligence. This leaves agencies to self-regulate, often using guidelines that are voluntary, opaque, or written in collaboration with the very companies that benefit from lax oversight.
Bloch-Wehba’s research highlights that the government’s reliance on commercial AI tools frequently evades constitutional scrutiny. For example, when a predictive policing tool influences patrol routes, the decisions are made by a private algorithm, not a public official. This creates a difficult legal question: does the Fourth Amendment apply to a search that is effectively instigated by a proprietary model? The answer, according to Bloch-Wehba, is far from clear, and courts have been reluctant to pierce the corporate veil to examine the logic embedded in these systems. As a result, AI regulation in the public sector is often a matter of corporate best practices rather than constitutional mandate.
The Police Surveillance Complex: Palantir and the Warehouse of Data
Few companies illustrate the fusion of tech and public power as vividly as Palantir. Originally founded with CIA venture capital, Palantir has become the go-to firm for law enforcement agencies seeking to aggregate vast amounts of data. Bloch-Wehba points to Palantir’s Gotham platform, which integrates everything from arrest records to social media activity into a single interface for investigators. The platform is not merely a tool; it is a conceptual framework that shapes how police understand and act on information.
The problem, as Bloch-Wehba articulates, is that these systems lack transparent rules. There are no clear guidelines on data retention, no public oversight of the algorithms that determine relevance, and no independent audit mechanism. When a police department subscribes to Palantir, it gets a black box that claims to identify connections, threats, and patterns. The burden of proof shifts, she argues, from the state to the citizen, who has no way to understand how they were flagged or why.
Moreover, the economic model supports this opaqueness. Palantir and similar firms do not sell simple software licenses; they sell ongoing contracts that include training, support, and customization. This creates a long-term dependency that makes it difficult for agencies to change vendors or adopt open-source alternatives. The result is a surveillance ecosystem where public safety is outsourced to a for-profit entity whose primary loyalty is to its shareholders, not the public interest.
Bloch-Wehba does not spare the Justice Department from criticism. She highlights that federal authorities have used these tools to justify decisions in immigration enforcement, counterterrorism, and domestic policing. The sheer scale of data collection—often justified as necessary to prevent crime—has been normalized to the point that it feels archaic to question it. But her work asks a fundamental question: if a police officer cannot explain why a search was conducted, is the search still lawful?
Why Did the Government Outsource Core Functions? The Efficiency Myth
The conventional narrative is that outsourcing to tech companies improves efficiency, reduces costs, and leverages private-sector innovation. Bloch-Wehba suggests this narrative is at best incomplete and at worst misleading. In practice, she observes, outsourcing has led to fragmented accountability, higher long-term costs, and a loss of institutional expertise within government agencies.
When a government agency contracts with a tech firm, it often loses the ability to understand the technology it depends on. This is a well-documented phenomenon across federal and state entities. Staff may have access to dashboards and reports, but they lack the engineering knowledge to inspect the underlying code, test for bias, or challenge the vendor’s conclusions. Bloch-Wehba describes this as a form of “intellectual surrender” to the private sector, which is particularly dangerous in areas like criminal justice where decisions have profound human consequences.
The legal dimensions of this outsourcing are the core of her recent work, including her forthcoming publications on trade secrecy and government use of AI. Here, she introduces a crucial concept: trade secrecy as a shield against accountability. Tech companies often refuse to disclose their algorithms, citing proprietary interests. This creates a legal void where neither the public nor the courts can fully assess the legitimacy of a government action rooted in an algorithmic recommendation. Bloch-Wehba argues that this is an untenable position for a constitutional democracy, where the citizenry must be able to understand and contest the bases of state power.
Trade Secrecy and the Evasion of Due Process
Trade secret law is designed to protect competitive advantage, but when applied to public-facing technologies, it collides with due process rights. If a defendant is denied parole based on an algorithmically generated risk score, due process traditionally requires that they be able to understand the evidence against them. Yet trade secrecy can prevent them from seeing how the score was calculated.
Bloch-Wehba contends that the government, not the private company, is responsible for ensuring that this does not happen. The state cannot pass the buck to its contractor when constitutional rights are at stake. Yet in practice, government agencies have been all too willing to accept trade secrecy protections to avoid difficult questions about their algorithms. This allows them to make consequential decisions without public scrutiny, a situation that Bloch-Wehba calls a serious erosion of the rule of law.
Amateur Surveillance and the Rising Role of Non-State Actors
In a surprising extension of her analysis, Bloch-Wehba explores how the trend toward “amateur surveillance” is compounding the problem. She is referring to the explosion of investigative work done not by professional law enforcement but by private individuals, amateur sleuths, and online communities using commercially available tools. These actors are not bound by constitutional search-and-seizure rules, and their actions can have serious legal consequences when they spiral into privacy violations or mistaken accusations.
The technological tools underlying this amateur surveillance, Bloch-Wehba notes, are similar to those used by police. Facial recognition apps, geolocation data brokers, and social media scraping platforms are readily accessible to the public. When these tools are used by private hands, courts have historically applied a different—often more permissive—standard. The result is a legal gray zone in which the same type of data that would require a warrant if collected by police can be gathered by any individual with a subscription service.
This raises pressing questions about the future of privacy law. If the state can encourage or tacitly accept amateur surveillance, it can circumvent the Fourth Amendment. Bloch-Wehba suggests that new legal frameworks are necessary to address this, but she is skeptical that the current Supreme Court or Congress will act quickly enough to keep pace with technological change.
The Regulatory Gap: Why the Government Cannot Oversee What It Cannot Understand
The central challenge in AI regulation is not a lack of will but a lack of capacity. Bloch-Wehba emphasizes that Congress and federal agencies like the Federal Trade Commission (FTC) are poorly equipped to audit complex algorithms. They rely on expert testimony, which tech companies can bury under technical jargon, and on self-reported compliance, which is often little more than a public relations exercise.
Her research advocates for independent testing and auditing of algorithmic systems. She points to models used in other countries, where public authorities have the power to inspect source code and test for bias before a system is deployed. The United States, she argues, lags far behind these standards, and the legal community has been too slow to demand change.
The legal profession itself shares some blame. Lawyers who represent government agencies often lack the technical literacy to question vendor claims. Meanwhile, corporate lawyers for tech companies are adept at using contracts, warranties, and liability waivers to insulate their clients from responsibility. The result is a marketplace where the legal framework actually incentivizes ignorance: agencies that ask fewer questions get cheaper contracts, and companies that disclose the least gain the most business.
Public Accountability in an Age of Machine-Learning Governance
Bloch-Wehba’s prescription is not to ban technology from government, which is neither realistic nor desirable, but to restore a healthy boundary between the public and private sectors. This means establishing clear rules for when and how agencies can adopt algorithmic tools. It means creating databases of government-contracted algorithms that are open to public review, rather than hidden behind procurement documents. And it means ensuring that courts have the technical expertise—or access to court-appointed experts—to assess claims about algorithmic accuracy and fairness.
She offers a practical analogy: we do not let a private company build a bridge and then refuse to inspect its safety records because the construction details are a trade secret. The same logic should apply to algorithmic infrastructure. If an AI tool is used to decide that someone should lose their public benefits, be subjected to enhanced surveillance, or be denied bail, the public is entitled to know how that tool works. Anything less is an abdication of democratic governance.
The conversation also touched on the role of political pressure. While government insiders may be aware of these issues, the incentive structure often works against reform. Politicians benefit from appearing tough on crime, and if a surveillance tool makes them look effective, they are unlikely to scrutinize its origins. Bloch-Wehba argues that public pressure, civic education, and legal advocacy are essential to counterbalance this inertia.
What History Tells Us: Recurring Patterns of Private Power
This is not the first time American governance has grappled with the concentration of private power. Bloch-Wehba draws historical parallels to eras when private entities controlled essential public functions—such as early policing, prison labor, and even the collection of taxes. In each case, democratic societies eventually realized that certain powers should not be delegated to actors outside the public sphere. The challenge today is that technology moves faster than legal precedent, and the consequences of failure are often invisible until a crisis occurs.
The Supreme Court’s recent decisions on administrative law have added another layer of complexity. A more conservative judiciary, suspicious of broad agency authority, may be inclined to defer to private contracts over public rulemaking. This could further entrench the role of companies like Palantir, Amazon, and Clearview AI in shaping government practice. As Bloch-Wehba puts it, the law is being written to protect innovation at the expense of public rights, and there is no organized counterweight strong enough to push back.
The Concrete Stakes: What Citizens Can Demand
The implications of this hidden grip are not theoretical. AI tools are already being used to prioritize child welfare investigations in several states, a practice that has been criticized for disproportionately flagging low-income families. Law enforcement agencies use predictive mapping to concentrate patrols in neighborhoods, which can lead to racial profiling and self-fulfilling cycles of surveillance. Immigration enforcement uses data analytics to identify potential violators, raising serious civil liberties concerns.
In each of these examples, the technology is not neutral. It is the product of choices made by engineers and executives who are not accountable to the public. The data used to train these systems often reflects historical biases, and the models can perpetuate those biases at scale. Bloch-Wehba calls this a “jurisdictional mismatch”—the scale of harm is public, but the power to redress it is private.
Where Do We Go from Here? A New Legal Realism
Bloch-Wehba is not without hope. She sees a growing awareness among scholars, journalists, and judges about the opacity of AI in government. Lawsuits challenging algorithmically driven decisions are beginning to succeed, forcing agencies to disclose more information. The introduction of algorithmic impact assessments in some cities, like New York’s law regulating automated employment decision tools, offers a template for other jurisdictions. However, these are small steps relative to the scale of the problem.
The next decade will be critical. If lawmakers and the legal profession continue to treat AI as a neutral tool requiring only technical fixes, the public will lose the ability to contest increasingly consequential state actions. If, however, the legal system adapts to treat algorithmic governance as a constitutional moment—one that demands transparency, discoverability, and oversight—then it is possible to harness the benefits of AI without losing the core of democratic accountability.
As Hannah Bloch-Wehba’s work demonstrates, the choice is not between technology and tradition. It is between a legal system that abdicates its responsibility and one that rises to meet the complexity of the digital age. The public should not accept the fiction that private tech companies are outside government, because they are already at its center. The only meaningful question left is whether the law will catch up to that reality in time. For judges, litigators, and policymakers willing to listen, her research offers not just a warning, but a practical guide for the hard work of reclaiming public power from the shadows of Silicon Valley.
