Important questions are raised currently about the stability of the relations between security intelligence agencies and their parent ‘liberal democratic’ governments. Populist electoral movements have already given rise to illiberal authoritarian nationalist governments in, for example, Argentina, Brazil (2019-23), Czechia, Hungary, India, Israel, Slovakia and the United States and lead in the polls in France and Germany.
It is not clear that all these governments have clashed with their intelligence agencies, but there are examples of this occurring. The most obvious is the United States where it was suggested that Gina Haspel, when appointed CIA Director in 2018 would be the first director who ever had to confront the problem of what to do when the president of the United States was a threat to national security because of his relationship with Vladimir Putin. [1]
Emboldened by his re-election in 2024, Trump appointed the inexperienced Tulsi Gabbard as the Director of National Intelligence (DNI), who then applied loyalty tests to potential recruits relating to their voting record and belief about the ‘stolen’ 2020 election. Gabbard fired the top two officials of the National Intelligence Council after their analysis challenged arguments that the Venezuelan government directs the Tren de Aragua gang, which had been Trump’s rationale for invoking the Alien Enemies Act. In 2025 Gabbard revoked the security clearances of 37 former and serving officials (effective dismissal for those still serving) where the common factor was their involvement in the 2017 assessment of Russia’s interference in the 2016 election. In the same year, the FBI forced out three senior officials who had either been involved in investigating the January 6, 2021, Capitol Hill riot or resisted White House efforts to identify other agents who were.
In Israel in April 2025, under pressure from PM Netanyahu, Ronen Bar announced that he would resign as Director of Shin Bet. The Supreme Court granted a temporary injunction and Bar submitted an affidavit to the Court (part public and part classified) in which he said Netanyahu demanded that he make false claims of security risks in order to extricate the PM from his corruption trial, that Bar obey him rather than the Supreme Court in the event of a constitutional crisis and that Bar take action against anti-government protesters. Also, in 2025 forty-one officers within the IDF Intelligence Directorate wrote to Netanyahu saying they would refuse to take any further part in the Gaza offensive, for example, selecting bombing targets.
In Germany the domestic security intelligence organisation Bundesamt für Verfassungsschutz (BfV) had already classed Alternative für Deutschland (AfD) as right-wing extremist and so incompatible with the free democratic order in three eastern states and in May 2025 this was extended nationally. This determination, which survived a court challenge from AfD, permits increased covert surveillance of AfD by informants and interception of communications etc. AfD were second in federal elections in February 2025 with 21% of the vote and have 152/630 seats in the Bundestag, therefore it is not unrealistic to imagine them as part of a future ruling coalition in Germany and the consequent potential for a clash between professionals and government.
The role of domestic agencies is to protect the regime against national security threats. It was only after 1945 that a few countries, mainly the victims of Nazi occupation, introduced legislation that provided a legal (rather than solely pragmatic) basis for agency actions and basic oversight structures. It was the 1970s before more liberal democratic countries followed suit, mainly in response to scandals of excessive surveillance of citizens by internal agencies. Following the end of the Cold War and the attempt to democratise Eastern Europe and the former Latin American military dictatorships, legislation mandating agency powers and oversight became widespread. These laws tended both to empower the agencies and to restrict them in certain areas but one key aim was to make security organisations more accountable to elected ministers. Paradoxically, this democratic principle now provokes the question of how agencies defend democratic principles that are under attack from elected authoritarian governments.
Democratisation continued into the new century but since 2008 has ground to a halt for several reasons: economic, reflecting the impact of the financial crash on incomes and social including the increasing fears around immigration both in the US and Europe. As a result, populist proponents of various forms of illiberal democracy have prospered and even if they have not won power their impact on governance has been significant. There is an extensive literature on what is described as ‘democratic backsliding’ in general but little analysis of its effect on security intelligence agencies. Perhaps this is because they are assumed to be such reliable bastions of support for governments whatever their policies, but it is the very centrality of the agencies to the survival of governments that requires specific consideration of how they deal with trends towards illiberal governance.
Any illiberal government depends on loyal security organs to stay in power, but we cannot assume that all the agencies involved in security governance act together or speak with one voice: ‘bureaucratic politics’ may rule. Technological changes in the twenty-first century have enhanced the agencies’ capacity for mass surveillance through their symbiotic relationship with the corporate suppliers of communications, Internet and social media in an overall structure that might be described as surveillance corporatism. While these agencies will be a tool of authoritarian governments, they may also be their victims. So, if constitutional checks and balances are being eroded through the actions of elected authoritarians, how will, or should, security intelligence agencies react?
On the face of it, the answer is simple: from an instrumental perspective bureaucrats act neutrally to implement the policies of the executive power, but authoritarian leaders view the bureaucracy as part of the ‘swamp’ and seek to change it into a loyal extension of their power. To the extent that these governments see themselves as opposed by varieties of ‘undesirables’ – socialists, eco-warriors, Islamists, migrants – leaders will define the agencies’ role in traditional ways: the surveillance and disruption of groups who may resist or take action, however peaceful, against the government of the day. As such, the agencies have more often been viewed as potential threats to liberal democracy.
In many countries no doubt this simple answer still pertains but we might consider an alternative institutionalist perspective in which bureaucrats are ‘guardians of state institutions and protectors of the democratic way of life’. [2] Officials cannot be value neutral and purely instrumental but are responsible for defending the principles and institutions of liberal democracy including constitutionalism, the rule of law and the public interest.
But, have the changes in law, governance, recruitment, training, working cultures and oversight of the past half century produced internal security agencies which will not simply do the bidding of ministers but will push back against them when they believe their requests/orders are illegal or unethical? Even if intelligence officials believe that a policy is mistaken or likely to be counterproductive, though not actually unconstitutional, they are, to quote the well-worn aphorism, obliged to ‘speak truth unto power’. In practice, that can be difficult, but how much stronger is the requirement if an executive proposal is seen as unconstitutional? Might the agencies become less the tools of authoritarians’ rule by law and more the defenders of rule of law?
It is possible to identify an escalatory ladder of resistance: ignoring demands, submission of critical reports, whistleblowing, active disruption and culminating in resignation which might all be legitimate if based on a proportionate response to the executive, but the serious difficulties and potential costs facing resisters are undeniable. Even if resisters identify what is to them an illegal use of executive power, it is likely to be characterised by authoritarians, not as legitimate defence of the institutional order, but as confirming their claim that they face the opposition of a ‘deep state’. As the earlier examples show, executives and their loyalist agency directors may simply dismiss resisters or take disciplinary action against them therefore, although resisters may find there is some protection to be had in group solidarity, resistance may cost them dear. But if officials see that the rule of law and accompanying liberal order are at stake is there not a duty to disobey?
Peter Gill
Visiting Professor
School of Criminology, Sociology and Policing
University of Hull
UK
pgill806@gmail.com
[1] Tim Weiner, The Mission: the CIA in the 21st century, Willaim Collins, 2025, p.313.
[2] Michael Bauer, ‘Administrative responses to democratic backsliding: When is bureaucratic resistance justified?’ Regulation and Governance, open access 2023 p.7 (18:4, 2024, 1104-1117); see also Cüneyt Gürer and Elena Walczak, ‘Democratic Backsliding and Security Governance, Connections, 23(4) 2024, pp9-31; Kutsal Yesilkagit et al, ‘The Guardian State: Strengthening the public service against democratic backsliding, Public Administration Review, 84, 2024, pp.414-25. Back to Table of Contents
