Federica Carugati is a lecturer in history and political economy at King’s College London.
In 2020, Vice revealed that the U.S. military had signed a contract with Babel Street, a Virginia-based company that created a product called Locate X, which collects location data from users across a variety of digital applications. Some of these apps are seemingly innocuous: one for following storms, a Muslim dating app and a level for DIY home repair. Less innocuously, these reports indicate that the U.S. government is outsourcing some of its counterterrorism and counterinsurgency information-gathering activities to a private company.
While states have always collected information about citizens and their activities, advances in digital technologies — including new kinds of data and infrastructure — have fundamentally altered their ability to access, gather and analyze information. Bargaining with and relying on non-state actors like private companies creates tradeoffs between a state’s effectiveness and legitimacy. Those tradeoffs might be unacceptable to citizens, undermining our very understanding of what states do and how we should interact with them.
Information And The State
The state, Max Weber famously noted, is “the only human community that (successfully) claims a monopoly on legitimate physical violence within a certain geographic territory.” Social scientists since then have identified three core functions of the state: coercion, extraction and administration. To extract, administer and coerce, the state needs information — to collect taxes, for example, or enforce laws or respond to potential threats.
Before the modern period, autocracies and democracies differed in the strategies they employed to gather information. Around 2000 BCE, for example, the Kingdom of Ur in present-day Iraq developed a standing army, a unified administrative system, a tax system, schools for bureaucrats and accounting procedures. Classical Athens (508-322 BCE) was a little different — it had no police force, centralized tax system or bureaucracy, so it relied on private citizens for those tasks.
Since the dawn of the modern period, the state has been able to obtain information through institutions it directly controlled, like centralized bureaucracies, professional police forces, formally trained officials (such as judges and public prosecutors) and permanent militaries. To rephrase Weber’s definition, then, the modern state is the only human community that successfully monopolizes information to perform core functions within a certain geographic territory.
Digital States
Digital technologies have, however, caused states to lose their monopoly over information. As a result, states are no longer the most effective or efficient actors at accessing, gathering and analyzing information.
This loss is driven in part by the nature of digital information. The enormous volume, speed and longevity of digital data require novel tools. States can choose not to develop such tools, with potentially dramatic implications for governance and security. Or they can try to develop them in-house, as many U.S. agencies are doing with AI algorithms. The resulting products, however, tend to lag behind those developed by private companies in terms of sophistication, and their use raises important institutional and ethical concerns.
States can also outsource the development and deployment of these tools, as the U.S. did with Babel Street. But that can lead to crises of legitimacy and public trust, especially when information-processing algorithms misfire. Palantir, for example, developed crime-forecasting software for police departments that was exposed for “techwash[ing] bias by making sloppy investigative work seem objective.”
The challenges that states face in gathering and analyzing information are also due to a loss of access. Information that the state needs might lie behind corporate walls or in the hands of citizens who resist releasing it, forcing states to buy or forcibly take it. China, for example, outsources repression to non-state groups, often young men with no formal military training but with more localized information about perceived threats to the state.
In the digital age, then, the state faces an age-old question about how to develop the capacity to access, gather and analyze information, but a new environment in which there is more information that is distributed across novel channels controlled by a variety of actors. Whether a state chooses to invest in such capacity in-house or forge relationships with independent actors will have profound consequences for its legitimacy and relationships with citizens and corporations. These choices also have ripple effects across other state functions.
Coercion
In 2015, Syed Rizwan Farook and Tashfeen Malik violently attacked a social services center in San Bernardino, California. Afterward, the FBI sought to access information on the attack from Farook’s iPhone. Apple refused, citing privacy concerns. A lengthy legal battle followed. Unable to reach an agreement with Apple, the FBI partnered with an Australian firm, Azimuth Security, which was ultimately able to gain access to the phone.
The encryption confrontation in the San Bernardino case illustrates the extent to which states might find themselves in the infelicitous position of petitioning private companies in order to access information relevant to national security — publicly failing to obtain it, and then having to pay someone else to do it.
The FBI was perhaps embarrassed, but the survival of the state itself was not at stake. That is not always the case. On the eve of the Arab Spring, Egypt’s then-President Hosni Mubarak shut down access to the internet in order to prevent coordination and communication among protesting citizens. To do so, he had to rely on private internet service providers. Mubarak is not alone. Iran has been working to build a national intranet with few points of access to the global internet, an effort to control all communication coming and going from the country.
Iran’s attempt at regaining an information monopoly in the realm of coercion illustrates a broader point: Negotiating with independent non-state actors may make a state less effective and legitimate in the eyes of its citizens, but having unchecked information capacity may also allow it to engage in repression much more frequently and easily.
Extraction
The state’s loss of the monopoly over information also matters for its ability to extract revenue. In the digital age, the mobility of information presents new challenges for monitoring economic transactions and following capital flows.
Because economic transactions, from online consumption to mobile banking to financial investments, have become visible digitally, taxation of these transactions may be more easily monitored and enforced, leading to increased revenue. For example, Ethiopia used to have notoriously low tax revenue, at times reaching just 6% of GDP, one of the lowest in the region. But since 2010, the country has been working to digitize its tax filing system, reducing the time required for filing and paying taxes and increasing monitoring and compliance. The result has been a jump of 12% in income taxes and 48% in value-added tax revenue.
However, it is easier now to hide assets or move them across sovereign borders to evade taxes. Private legal entities and multinational banks can hold information about taxable assets and their location, which creates multiple and conflicting layers of jurisdiction. According to the International Monetary Fund, between $500-$600 billion in state revenue is lost to tax evasion every year.
Recognizing the challenges that states face to enhance their own capacity and cooperate with non-state actors to better track and trace these assets, Janet Yellen (then the president of the IMF) proposed a global 15% corporate tax rate meant to mitigate the attractiveness of tax havens and prevent a race to the bottom in corporate tax evasion. However, the effectiveness of such a policy is undermined by the fact that crucial information — for example, about the nature of assets and transactions — may lie beyond any single state’s reach.
Administration
Digitization once held the promise of improving analog bureaucracies to create more transparent and equitable access to the state and its resources. In some places and for some purposes, the promise was fulfilled. More often, headline-catching failures and competition from private actors have eroded the state’s legitimacy.
Estonia was an early technology adopter, fully digitizing its records in the 1990s under the leadership of then-president Toomas Hendrik Ilves. Estonians can vote from home and all their data — from addresses to financial information to medical records — is collected and protected through digital IDs, cryptography and now blockchain technology.
This digitization has improved access to and delivery of health care and other government services in the country. Estonia’s small size and first-mover advantage facilitated the transformation, which remains but a mirage in many larger countries.
In the U.S., full digitization would be much more difficult to achieve, but some government agencies are capitalizing on digital technology to carry out their jobs more effectively. For example, the Department of Energy now houses the National Renewable Energy Laboratory, which maps the most effective locations for solar panels. At the local level, cities and local governments have invested heavily in developing information systems and GIS capacity to make better decisions about, for example, land use within their municipal boundaries.
But there have also been dramatic negative consequences to digitization, perhaps most visibly in the context of providing public goods. Virginia Eubanks, a writer and academic at SUNY Albany, documented Indiana’s experiment with automating welfare eligibility, for example, in which humans were replaced in the decision-making loop by algorithmic systems that ultimately limited access to vital services for vulnerable citizens for opaque reasons.
Indiana is hardly unique. To make up for the state’s failures, nonprofit and philanthropic actors have begun engaging or competing with the state to provide services or facilitate access to existing ones. That the state is failing at its job of providing public goods is the inspiration behind Code for America, whose armies of young hackers aim to make government more accessible and service delivery more equitable. Yet, as Robert Reich and Anand Giridharadas note, private philanthropy in public goods provision is both morally problematic and dangerous for democracy because it undermines incentives to hold leaders accountable.
Judicial systems are evolving similarly. Gillian Hadfield, a professor of law at the University of Toronto, has documented how multinational corporations are able to eschew state laws and regulations by building parallel judicial systems in-house. Other private actors, such as SquareTrade, a warranty provider owned by Allstate, are building judicial platforms online. These platforms are designed to facilitate access to justice and resolve disputes — for example between an eBay seller and an upset buyer — by avoiding allegedly slow and expensive public courts.
Proprietary software may speed up service delivery, but it can also delegitimize institutions and simultaneously facilitate inefficient and unfair outcomes in private systems that use biased data. In this respect, Northpointe Inc.’s algorithm to predict recidivism — COMPAS — is but one instantiation of an otherwise common pattern in which America’s racial past creeps into the future under the guise of equity-enhancing technology.
In the digital age, some states will be better at building information-gathering and analytical tools in-house; others will more successfully bargain for the information they need. As ancient states did, so modern digital states will choose between incentivizing independent actors and coopting agents they can control. The winning strategy, however, is not obvious.
Moreover, the digital state’s loss of the monopoly over information means that functions formerly associated with the state now exist in hybrid spaces dominated by rather obscure — unclear and therefore ultimately unaccountable — relations between private and public institutions.
Evgeny Morozov has suggested that algorithmic regulation signals the rise of data and the death of politics. But the strategic choices states must make about developing or outsourcing information capacity have consequences not just for state capacity, but also for non-state actors, including private citizens. The politics of the digital state, then, are alive and kicking. At stake is the legitimacy of governments in the digital age, as well as the power relations between them and those on whom they rely for information.
