Intelligence Oversight Research Articles

Information Standards and Guidelines on AI: Ethical Concerns, Trustworthiness, Quality Assessment, and Human Oversight

ABSTRACTThis panel, sponsored by the ASIS&T Standards Committee, has two purposes. First, the panel gives a brief report on the most recent activities of the national and international standards in the field of information science and technologies in which the ASIS&T Standards Committee and members have been involved. More importantly, the panel will focus on several recent significant standardization efforts that address the ethical concerns, trustworthiness, quality assessment, and human oversight of artificial intelligence (AI) systems. A final open discussion on these issues will involve the audience, to highlight the benefits, challenges, and standards needed while addressing the important role of information science within our evolving society.

Intelligence Oversight and Accountability in Poland: Has Anything Changed?

The years 1990 and 1991 marked the beginning of a sea change in this particular area of the functioning of the Polish state. The following phenomena were analyzed: Over three decades on from the restitution of the Polish democratic state, how have the focal issues of key intelligence oversight institutions been addressed? How successful was the political transition in the area of intelligence services? The article is composed of four main parts. The first part provides a discussion of the “special services” concept, which determines the Polish institutional framework for intelligence oversight and accountability. Subsequent sections focus on particular forms of intelligence oversight: executive control, parliamentary oversight, and independent oversight. Our focus is on the institutional dimension of state security. The article is intended as an overview of the issue and relies on traditional historical and institutionalist approaches. Our proposal is to conduct a detailed case study to present the current situation in a country in East Central Europe, with the aim of generating general findings that could be used to guide reform efforts.

Book Review – Intelligence Overview: The Watchers and the Watchdog

Book Review by Dr Peter Grace Didier Bigo, Emma Mc Cluskey, Felix Tréguer (eds), (2024), Intelligence Oversight in Times of Transnational Impunity. Who will Watch the Watchers? Oxon: Routledge. ISBN 9781032406541. 310 pages. This book focuses on two important issues in intelligence oversight: the long-standing acceptance of raison d’état, and the increasing relations between intelligence agencies transnationally and how much those should come under scrutiny and legal restraint. It is not a book that is designed to be read comfortably by intelligence practitioners, and indeed there are priors that suggest practitioners would much prefer it if there was no oversight of their activities at all. Whether framing the enquiry in this way provokes some degree of urgency or simply divides opinion, the questions of raison d’état and transnational oversight are important.

A blue ribbon goat: the Rockefeller Commission, public opinion, and the ford administration’s intelligence reform failure

ABSTRACT After allegations of the US intelligence community conducting illegal operations that targeted American citizens, 1975 became known as the Year of Intelligence. The subject dominated national discourse and led to competing programmes of intelligence oversight proposed by Congress and the executive branch. This article examines the Ford administration’s programme that started with the Rockefeller Commission in January 1975 and ended with the issue of Executive order 11,905 in 1976. It argues that neither were enough to stave off Congress intruding on executive control of the intelligence community, because the Ford administration inaccurately measured the public’s demands for genuine reform.

When Oversight Went Awry: Congress and Intelligence in the Twenty-First Century

Congress’ intelligence oversight has gone off the rails. The legislative branch has taken its constitutional responsibilities of surveying and guiding national intelligence and instead practiced rampant partisanship. The formerly bipartisan and sober intelligence committees have become a political battleground for most of the twenty-first century, although each body will occasionally exhibit rationality and responsibility. Congress began by treating intelligence as another issue to be addressed in the context of contemporary politics. In the Trump era, intelligence itself became a central battleground, and the oversight structure was the forum for a presidential impeachment. Seen in three phases, intelligence oversight in Congress has gone from partisanship (2003–2004) to hyperpartisanship (2004–2014) to intermittent dysfunction (2017–present). This assessment considers Congress’ intelligence oversight through numerous case studies and offers positive examples for reform.

Overt action: congressional oversight, private activism and Afghan covert action policy in the Reagan administration

ABSTRACT Intelligence scholars routinely portray intelligence oversight as a means of restraining intelligence activities, particularly covert operations. The consensus overlooks situations where oversight can instead channel popular passions. This paper documents how Reagan era activists recruited legislators to the Afghan resistance cause. Their legislative-civic alliance demanded the CIA launch more aggressive operations in Afghanistan, sidelining established oversight committees. The resulting covert action campaign risked Soviet escalation, eliminated plausible deniability, and gave advanced U.S. technology to potential terrorists. This episode highlights how well-organized lobbyists may affect the intelligence agenda and challenges assumptions that wider engagement in oversight will always restrain intelligence agencies’ overreach.

Institutionalised distrust and human oversight of artificial intelligence: towards a democratic design of AI governance under the European Union AI Act

Abstract Human oversight has become a key mechanism for the governance of artificial intelligence (“AI”). Human overseers are supposed to increase the accuracy and safety of AI systems, uphold human values, and build trust in the technology. Empirical research suggests, however, that humans are not reliable in fulfilling their oversight tasks. They may be lacking in competence or be harmfully incentivised. This creates a challenge for human oversight to be effective. In addressing this challenge, this article aims to make three contributions. First, it surveys the emerging laws of oversight, most importantly the European Union’s Artificial Intelligence Act (“AIA”). It will be shown that while the AIA is concerned with the competence of human overseers, it does not provide much guidance on how to achieve effective oversight and leaves oversight obligations for AI developers underdefined. Second, this article presents a novel taxonomy of human oversight roles, differentiated along whether human intervention is constitutive to, or corrective of a decision made or supported by an AI. The taxonomy allows to propose suggestions for improving effectiveness tailored to the type of oversight in question. Third, drawing on scholarship within democratic theory, this article formulates six normative principles which institutionalise distrust in human oversight of AI. The institutionalisation of distrust has historically been practised in democratic governance. Applied for the first time to AI governance, the principles anticipate the fallibility of human overseers and seek to mitigate them at the level of institutional design. They aim to directly increase the trustworthiness of human oversight and to indirectly inspire well-placed trust in AI governance.

Partisanship and congressional intelligence oversight: the case of the Russia inquiries, 2017-2020

ABSTRACT What patterns of behavior can we observe among the leading figures of US congressional intelligence committees’ Russia inquiries? We argue that current typologies to study the roles of legislators in intelligence oversight do not fully capture the impact of partisanship on overseers’ behavior. Revisiting Loch Johnson’s typology of ‘role definitions’ to account for the behaviour of these legislators, we define four types of congressional investigators to understand how partisanship affected legislators’ behavior during the congressional investigations on Russian interference. We conclude by showing how our role definitions could be used to study congressional intelligence oversight on other high-profile issues.

Civic Intelligence Oversight: Practitioners’ Perspectives in France, Germany, and the UK

In recent years, various revelations about government malfeasances have highlighted the vulnerability of civil society actors who work on surveillance by intelligence agencies. Simultaneously, new technologies and overburdened state oversight bodies clarify how relevant citizen scrutiny of intelligence is. Both of these factors have led to the emergence of scrutiny by civil society actors as a research subject. This paper contributes to such scholarship by presenting data collected through surveys addressed at journalists and professionals from civil society organisations (CSOs) in France, Germany, and the UK to comparatively characterize the forms, scope, and constraints of the scrutiny they perform. Indicated differences across countries highlight variances in the practices of civic intelligence oversight. These variances indicate that there is room to manoeuvre for civic forms of holding intelligence agencies to account, counteracting the primacy of security and the secrecy of intelligence. Yet, similarities of civic oversight practitioners’ perspectives across all three countries are also distinct and informative; in particular, across all three countries, journalists and CSO professionals who work on surveillance by intelligence agencies worry they are under surveillance themselves and express dissatisfaction with safeguards at work.

Assessing intelligence oversight: the case of Sweden

ABSTRACT The study of intelligence oversight captures the inherently political nature of secret intelligence. However, many studies of intelligence oversight adopt rather instrumentalist views that omit important political aspects of the policy process. Typically, these studies focus on obstacles to effective oversight. This article discusses how the effectiveness of oversight can be assessed by applying broad evaluative categories that contain programmatic, process-related, political, and durability dimensions. Empirically, the study probes the case of Sweden as an illustration. Swedish oversight arrangements have on balance been successful in some dimensions, particularly the programmatic dimension, which arguably also contributed to its relative longevity.

More Than Just Blind Guardians?: A Legal Analysis of Finnish Parliamentary Oversight of Intelligence

More Than Just Blind Guardians?: A Legal Analysis of Finnish Parliamentary Oversight of Intelligence

Parliamentary Intelligence Oversight in Romania: Between Consolidation and Controversy

Romania represents one of the paradigmatic cases where security sector reform proceeded quickly after the collapse of the communist regime in 1989. The former Securitate was one of the most repressive intelligence services in the communist bloc and drastic reform was necessary in order to establish an efficient and trusted domestic intelligence service. Previous research by Bruneau and Matei highlighted the important steps that Romania has undertaken over the past years and argued that an efficient and robust oversight system has been built. This article addresses contemporary debates on factors triggering reform in parliamentary intelligence and focuses on Romania’s 2017 reforms. These reforms can be analyzed through the application of the “fire alarm/police patrol” model developed by Loch K. Johnson for changes in congressional oversight. By increasing and clarifying the power of the Joint Standing Committee for the Oversight of the Romanian Intelligence Service, the 2017 changes brought more clarity and more accountability to the oversight system. After 2017, the committee reverted to a “police patrol model” focusing on visits to intelligence sites and meetings with high-level intelligence officials.

Towards democratic intelligence oversight: Limits, practices, struggles

AbstractDespite its common usage, the meaning of ‘democratic’ in democratic intelligence oversight has rarely been spelled out. In this article, we situate questions regarding intelligence oversight within broader debates about the meanings and practices of democracy. We argue that the literature on intelligence oversight has tended to implicitly or explicitly follow liberal and technocratic ideas of democracy, which have limited the understanding of oversight both in academia and in practice. Thus, oversight is mostly understood as an expert, institutional and partially exclusive arrangement that is supposed to strike a balance between individual freedom and collective security, with the goal of establishing the legitimacy of and trust in intelligence work in a national setting. ‘Healthy’ or ‘efficient’ democratic oversight then becomes a matter of technical expertise, non-partisanship, and the ability to guard secrets. By analysing three moments of struggle around what counts as intelligence oversight across Germany, the UK, and the US, this article elucidates their democratic stakes. Through a practice-based approach, we argue that oversight takes much more agonistic, contentious, transnational, and public forms. However, these democratic practices reconfiguring oversight remain contested or contained by dominant views on what constitutes legitimate and effective intelligence oversight.

Twenty years on: Intelligence and Security Committee and investigating torture in the 'war on terror'

ABSTRACT The central objective of democratic governance of intelligence is, through debate and law, to establish public confidence that the agencies work efficiently, effectively and properly. Oversight of intelligence can be seen as a contest between agencies, government and overseers for the control of information. The four interacting dimensions of information control are secrecy, gathering, evaluation and persuasion. This article assesses the oversight performance of the Intelligence and Security Committee (ISC) through the prism of information control in investigating the allegations of UK involvement in torture since 2001. Operating within an overall context of executive dominance, these dimensions constitute a series of filters including what officers tell their managers, what the agencies record, what they tell ministers, what they tell oversight bodies and, finally, what the ISC reports to the public.

Institutionalised Distrust and Human Oversight of Artificial Intelligence: Toward a Democratic Design of AI Governance under the European Union AI Act

Institutionalised Distrust and Human Oversight of Artificial Intelligence: Toward a Democratic Design of AI Governance under the European Union AI Act

Regulatory Capture of Intelligence Oversight Committees: A New Method Applied to the Polish Case

Regulatory Capture of Intelligence Oversight Committees: A New Method Applied to the Polish Case

Australian intelligence oversight and accountability: efficacy and contemporary challenges

ABSTRACT In 2017, the results of a major independent intelligence review have sparked transformational changes in Australia’s National Intelligence Community (NIC). Much of this change is still underway, but less thought has been given to how reform changes to both national security and home affairs intelligence capabilities might affect existing oversight and accountability mechanisms. This article will assess how traditional oversight and accountability mechanisms of the NIC may be affected by these transformational changes and whether current mechanisms are likely to remain fit for purpose.

The oversight of autonomous artificial intelligence: lessons from nurse practitioners as physician extenders.

The development of autonomous artificial intelligence (A-AI) products in health care raises novel regulatory challenges, including how to ensure their safety and efficacy in real-world settings. Supplementing a device-centered regulatory scheme with a regulatory scheme that considers A-AI products as a ‘physician extender’ may improve the real-world monitoring of these technologies and produce other benefits, such as increased access to the services offered by these products. In this article, we review the three approaches to the oversight of nurse practitioners, one type of physician extender, in the USA and extrapolate these approaches to produce a framework for the oversight of A-AI products. Under the framework, the US Food and Drug Administration would evaluate A-AI products and determine whether they are allowed to operate independently of physician oversight; required to operate under some physician oversight via a ‘collaborative protocol’ model; or required to operate under direct physician oversight via a ‘supervisory protocol’ model.

İSTİHBARATIN DENETİMİ: KARŞILAŞTIRMALI BİR ANALİZ

İstihbarat örgütlerinin denetimi tehdit parametreleri ve içinde bulundukları kurumsal kültürün yanı sıra ülkelerinin rejim türüne göre de şekillenmektedir. Bu bağlamda, bu makalede, kapsamlı bir literatür taraması ve ilgili açık kaynak istihbarat dokümanlarının incelenmesi gerçekleştirilerek demokrasiler ve otoriter rejimlerde gerçekleştirilen istihbaratın denetimi karşılaştırmalı olarak analiz edilmiştir. Yapılan çalışma sonucunda demokratik ve otoriter rejimlerde istihbaratın denetiminin temel olarak iki noktada farklılık gösterdiği tespit edilmiştir. Bunlar amaç farklılığı ve denetim mekanizması farklılığıdır. Amaç farklılığı noktasında demokratik ve otoriter rejimler arasındaki en temel ayrımı yasalara uygunluk denetimi ve toplum/rejim güvenliği amacı oluşturduğu sonucuna varılmıştır. Demokratik ülkelerde istihbarat denetimi, gerçekleştirilen faaliyetlerin kişisel haklar, ulusal ve uluslararası hukuka uygunluğunun incelenmesini içermekteyken, bu durumun otoriter rejimlerde sembolik bir süreci ifade ettiği tespit edilmiştir. Demokratik ve otoriter rejimlerdeki amaç farklılığından kaynaklanan pratiklerin kurumsallaşması sonucu denetim mekanizmalarının etkinliğinin de farklılaştığı sonucuna varılmıştır. Bu bağlamda, günümüzde demokrasilerin daha etkin ve kurumsal istihbarat denetim mekanizmalarına sahip olduğu tespit edilmiştir. Aradaki farkı oluşturan unsurların ise etkin denetim yetkisi ve sivil kontrol ile uzmanlaşma olduğu belirlenmiştir.

MT6 Oversight of Artificial Intelligence in Medicine: A Review of Frameworks

Artificial intelligence (AI) is rapidly expanding in medicine even while lacking formal oversight. We sought to identify and describe considerations for the oversight of AI in medicine. We also explored where along the translational process (i.e. AI development, reporting, evaluation, implementation, and surveillance) these considerations were targeted.