Digital Warfare: How Officials Bankrupted 10,000 Families

Third Party Mechanic (3PM)
11 min readJan 26, 2021


Welfare state authorities in the Netherlands combined innovations in artificial intelligence (AI) and regulatory powers, not merely for surveillance, but also to frame citizens for financial crimes. Can it happen here in the U.S.? The IRS has extensive prosecutorial powers and a history of targeting citizen groups. How closely should we examine their police function? Photo by h heyerlein. Follow @ heyerlein.

On Jan. 15, 2021, the Dutch Prime Minister announced his resignation shortly after investigators disclosed that the nation’s tax bureau had falsely accused over 10,000 families of fraud over child subsidies with the intent to force them into financial ruin. His entire ministry then simply walked out.

The fraudulent enterprise centered primarily on families of ethnic origin. Under the auspices of eradicating fraud by social welfare recipients, the Dutch government activated innovations in its artificial intelligence arsenal to collect and manipulate data that citizens had filed with their benefits paperwork.

When considering whether systematic abuse of digital power could be levered against US citizens in like manner (for example by the IRS), we first had to ask ourselves how anyone would find out? If there were similar misconduct or patterns of abuse within the infrastructure of the Internal Revenue Service (IRS), would we even notice?

SEE: Dutch Government Quits Over “Colossal Stain” of Tax Subsidy Scandal, Reuters (Jan. 15, 2021)

EXCERPT: Prime Minister Mark Rutte’s government resigned on Friday, accepting responsibility for wrongful accusations of fraud by the tax authorities that drove thousands of families to financial ruin, often on the basis of ethnicity. . . .

A parliamentary inquiry found last month that officials at the tax service had wrongly accused families of fraud over childcare subsidies, causing an “unprecedented injustice”. Around 10,000 families had been forced to repay tens of thousands of euros each, in some cases leading to unemployment, bankruptcies and divorces. Many of the families were targeted based on their ethnic origin or dual nationalities, the tax office said last year. . . .

“This is about tens of thousands of parents who were crushed under the wheels of the state,” Rutte told journalists. “There can be no doubt, this is a colossal stain.”

What makes the Dutch example far more chilling is actually omitted from early news reports in the United States about the historic resignations. Specifically, the Dutch government had engaged extensive data collection operations using artificial intelligence (AI) systems, called “SyRI.” The ostensible intent was to track fraud by social welfare recipients, but it essentially consisted of investigative strike teams, as the following court report details:

SEE: Netherlands: Court Prohibits Government’s Use of AI Software to Detect Welfare Fraud, Library of Congress: Law Library Blog (March 13, 2020)

EXCERPT: On February 5, 2020, the District Court of The Hague . . . held that the System Risk Indication (SyRI) algorithm system, a legal instrument that the Dutch government uses to detect fraud in areas such as benefits, allowances, and taxes, violates article 8 of the European Convention on Human Rights (ECHR) (right to respect for private and family life).

The [Dutch] SyRI system is an algorithm used by the government “to prevent and combat fraud in the fields of social security and income-related schemes, tax and social insurance contributions, and labor laws.” It links and analyzes data from various government or public agencies and generates a risk report if a person is suspected of fraud. In 2003, these agencies, which include, among others, municipalities, the Employee Insurance Agency (UWV), the Social Insurance Bank (SVB), the national tax authorities, the Immigration and Naturalization Service (IND), and the Inspectorate SZW, concluded a Cooperation Agreement and formed a National Steering Committee on Intervention Teams (LSI).

Therefore, according to the above cited records in the custody of the US Library of Congress, the Dutch government deployed tech warfare on unsuspecting citizens. They perverted the regulatory function to target innocent families who were mandated by law to provide the government sensitive information in the first instance.

Is reliance on the media or government officials to inform us about potential risks to our privacy and other freedoms well placed? If so, how do we explain the omission of any discussion regarding whether such AI tracking methods apply to the U.S. welfare state? We need to find out for ourselves. Photo by Mark Paton. Follow @ mark0polo.

Why were these facts about the role of AI not disclosed in U.S. reporting to date about the Dutch resignations? To the extent that U.S. legislators and the public rely on media reports to keep abreast of risks to democratic rights, we have yet to be alerted that a preventative review of digital welfare and tax administration is well warranted.

Here are some additional facts that may impact the breadth of the American public’s evaluation of whether there is a risk of the Dutch example replicating in the U.S.

Dutch welfare state authorities used their financial data mining systems to sweep entire neighborhoods.

“Nineteen of . . . 22 projects used a ‘neighborhood-oriented approach,’ meaning they targeted specific neighborhoods in which the linked data indicated an increased risk of welfare fraud.”

On Feb. 5, 2020, the District Court of the Hague ruled that “the development of new technologies means that the right to respect for private life, which includes the right to the protection of personal data, is increasingly important and that the absence of sufficient and transparent protection of it might have a ‘chilling effect’ among the population.. . . The Court [further] stated that people whose data is collected and included in a risk report are not automatically informed. There is only a legal requirement to announce the start of a SyRI project.” However, the Hague’s District Court stopped short of deciding whether the SyRI system was “justified” in light of the government’s need to detect fraud within social welfare programs because “it cannot determine what exactly the SyRI is.” The Dutch government had refused to provide its risk model to the Hague.

UN special rapporteur on extreme poverty and human rights, Philip Alston, submitted an amicus brief to challenge the Dutch program, noting that “welfare state authorities have made use of digital innovations since at least the 1960s. . ., but in recent years there appears to be a renewed and accelerated push toward digital government. The abundance of digital data available today, combined with rapid advances in computer power, storage and related innovations in the fields of data analytics and artificial intelligence have created a new impetus for digital government and digital welfare states.”

The Dutch government may well have resigned because of the clear understanding that, especially as the welfare state endures one of the greatest expansions since the Great Depression internationally, a program that selectively targeted minorities may now implicate broader communities of recipients. Everyone may be affected.

Turning to the American context, almost gone are the days when one could survive “off the grid.” With the advent of the global pandemic, economic circumstances have compelled relatively immense numbers of American citizens to seek financial support from state and federal government agencies. They have had to disclose all manners of personal information regarding their finances, as well as create tech accounts within the government’s infrastructure. Federal programs have further secured detailed information from small businesses and commercial entities, and disbursed payments through preexisting IRS channels. Moreover, as every locality gears to deliver vaccinations across the nation, online registration requirements are integral to effective public health administration. As innovations in government AI expand to absorb and manage the increased volumes, so too should citizen protections.

See Feb. 18, 2021 Update Note Below: Deloitte launches AI platform for U.S. government agencies.

As discussed below, can we harbor any reasonable assurances that the United State’s government, under the auspice of reducing benefit fraud, does not possess nor deploy such stratagems upon its citizenry when digesting the massive amounts of data it collects for social welfare programs — neighborhood by neighborhood?

Who Monitors the “Digital Welfare State”?

Perhaps all the recent necessary focus on regulating Big Tech companies has unwittingly diverted us from the reality that the biggest digital collecting agents in our lives are our own government’s bureaucracies. Photo by ev.

In the U.S., everyone over the age of 65 is eligible for social security. Any recipient of state subsidies or payer to the government is required to disclose a broad array of personal financial information, some of which we may actually share with no one else — not even a marital spouse. Anyone who has completed state applications for subsidies may well be able to recount the level of specific detailed disclosures required (e.g., unemployment benefits, food subsidies, school vouchers, etc.). With the Covid-19 pandemic, the number of those requiring government support as livelihoods have disappeared has grown exponentially in every state. Exactly how much personal financial data has the American public transferred to the state?

As the state apparatus aspires to tighten budgets (especially in the wake of national industry shut-downs), imposes stricter requirements on social benefit recipients, and enhances tracking technologies, advancements in digital technology is a pro forma method now for government officials to pursue an array of distinctly political objectives. In his amicus brief to the Hague, Philip Alston rightly points out that, “despite the increased attention given to the use of digital technologies in other areas of government (notably in the military, intelligence, policing, and immigration), there is still relatively little attention to digital innovation in social protection systems.

Simply, several agencies within the US federal system interface with citizens routinely, but also remain intensely obscure. One such behemoth would be the IRS. For example, have you ever heard of the “National Taxpayer Advocate”? That is the official at the IRS who is charged with “flag[ging] problems that can potentially make life difficult for people paying their taxes or who are trying to resolve a problem with their taxes. The advocate has previously looked at lengthy phone wait times, potential barriers for free filers and problems with the IRS’ own equipment and technology.” Still never heard of that post, although you have paid taxes annually since your first job? Neither had we.

SEE: There’s a New Watchdog for Taxpayers at the IRS, MarketWatch (Feb. 29, 2020).

There is also an Inspector General at the IRS, tasked with regular reviews of agency operations. That is who investigated Lois Lerner in 2013. Lerner was Director of the Exempt Organizations Unit of the IRS. In 2013, she was caught targeting conservative groups, such as the Tea Party, by either denying them tax-exempt status or delaying their applications to preclude them from contributing to their causes and candidates during the 2012 election.

SEE: Remember the IRS Targeting Scandal? No One Ever Got Punished for It, Washington Examiner (Jan. 18, 2018)

EXCERPT: Remember the scandal that erupted in 2013 when the IRS was found to be targeting conservative organizations for harassment? President Barack Obama called such conduct “outrageous.” “There’s no place for it,” he added, and assured the public that “[the IRS] have to be held fully accountable.”

And then nothing happened.

Finally, in October [2017], the IRS signed a consent decree in federal court in which it admitted to targeting conservative organizations for more than two years, from 2010 through 2013.

Yet, let us not readily forget that the Inspector General’s Office was also responsible for failing to secure Lerner’s emails, which conveniently disappeared when investigative subpoenas landed at the IRS.

SEE: IRS Lost Lois Lerner’s Emails in Tea Party Probe, CBS News (June 13, 2014)

EXCERPT: The IRS was able to generate 24,000 Lerner emails from the 2009 to 2011 because Lerner had copied in other IRS employees. The agency said it pieced together the emails from the computers of 83 other IRS employees.

But an untold number are gone. Camp’s office said the missing emails are mainly ones to and from people outside the IRS, “such as the White House, Treasury, Department of Justice, FEC, or Democrat offices.”

Certainly, there exist citizen groups that choose to mind a watchful eye over activities at the US tax collecting agency. However, they rely on private funding and do not have any special access into the operations of the IRS other than public information.

SEE: Citizens Against Government Waste

By comparison, the media reports that some of the targeted Dutch families are pursuing justice independently and at their own expense. Specifically:

In a rather unique move in the Netherlands, the families . . . filed criminal charges against five politicians, including Finance Minister Wopke Hoekstra and Economy Minister Eric Wiebes, for their role in the affair. If convicted, they could face up to six months in prison.

“We pressed criminal charges because the victims I represent have been ruined, some became homeless because of these policies. These politicians have been extremely negligent,” said Vasco Groeneveld, a lawyer representing 20 victims. “Every time I open their files, shivers run down my spine. These people have been treated awfully.”

SEE: Government in Netherlands Resigns After Benefit Scandal, New York Times (Jan. 15, 2021).

Still, it is unclear that such corruption could even be prosecuted under the criminal codes of the United States, especially without access to internal evidence of criminal intent and given various legal immunities that exist for discretionary state action. The anti-fraud AI program was an investigative law enforcement function. Even the Hague determined that Denmark had a compelling state interest in creating the data program, but had fallen with its execution. Come to think of it, the ambiguity of which government misconduct is actionable under U.S. codes of justice is itself disturbing.

Perhaps legislators and other officials within the US government will take note of the Dutch government’s example and inquire into whether there are comparable IRS protocols or systems levied against the populace. But, it is distressing that we are left to rely on the same bureaucratic structure that may be involved in such endeavors to identify, disclose, and end them. Is there another way?

Democrat Party officials seek to have Republican Party citizen groups investigated by the IRS following the Jan. 2021 Capitol riots, which is separate from any ongoing inquiries by the FBI or national security agencies. Photo by Andy Feliciotti. Follow @ someguy.

UPDATE NOTE (Jan. 25, 2021):

The IRS exerts state power well beyond the mere bureaucratic administration of personal or business finances. It also serves as gatekeeper for civic organizations to receive 501(c) non-taxable or charitable designations, the lifeblood for the existence and activities of many, including political speech. In the aftermath of the Jan. 2021 Capitol riots, prominent Democrat Party legislators are calling for investigations into whether several nonprofit organizations (student groups, women’s groups, etc.) that organized the original political rally should retain their tax-exempt status. It is unclear what protections, if any, targeted groups would receive in such an investigation, or whether their best defense would be private counsel — which is expensive.

SEE: A Top Senate Democrat Calls for IRS Probe of Dark-Money Groups that Helped Organize Rally Before Riot, CNBC (Jan. 15, 2021).

UPDATE NOTE (Feb. 18, 2021):

Jan. 2021 press release from Deloitte Consulting LLP advised that it released a new AI platform for government clients, expanding data mining capabilities for U.S. bureaucratic agencies. The press release did not indicate, however, what measures, if any, are built into the system to prevent government abuse comparable to the historic events in the Netherlands.

SEE: Deloitte Launches CortexAI for Government Platform, Modern Diplomacy (Jan. 30, 2021).

EXCERPTS: As the new Administration and statehouses across the country work to tackle key priorities across the dimensions of health, economic recovery, climate and racial inequities, data and its effective use through emerging AI technologies will be key enablers. To accelerate the adoption of AI in government, Deloitte today announced its new AI platform, CortexAI™ for Government. Designed to accelerate AI deployment among public sector organizations, CortexAI for Government combines proven government and industry solutions with AI models, tools and technical architectures tailored to support public sector missions with a focus on AI quality, audit and ethics.

“AI is more than technology. Through CortexAI for Government, we will help agency leaders and frontline public servants introduce new AI solutions and execute their organizations’ missions more effectively and with reduced cost,” said Ed Van Buren, principal with Deloitte Consulting LLP and leader of Deloitte’s public sector AI practice. “We are also deeply committed to advancing the ethical use of AI so that its use in American government becomes a model for the commercial sector and the world.”

Post originally published on, Third Party Mechanic (Jan. 15, 2021)



Third Party Mechanic (3PM)

INVESTIGATIVE, ILLUSTRATED & AFFORDABLE PUBLISHING about government influence in citizens’ lives. We ask. We tell. We push. WWW.3PMONLINE.COM