Fraud During a Pandemic: Identifying and Appraising New Challenges for the Criminal Justice Response in England and Wales
Lead Research Organisation:
University of Bristol
Department Name: Law
Abstract
The global Covid-19 pandemic has presented new opportunities for those looking to commit fraud, capitalising on new vulnerabilities (of systems, organisations, and members of the population), propelled by sophisticated cyber-enabled schemes. There has been an increase in fraud during the pandemic, as well as an increase in its complexity, which places it front-forward as a criminal law issue which requires advanced analysis. For example, on 7 May 2020, Action Fraud reported that a total of £2,996,252 has been lost by 1467 victims of coronavirus-related scams. 6069 reports of coronavirus-related phishing e-mails were received. Penalising fraud, especially against vulnerable members of the population such as older persons, has been held to be an 'immediate' charging priority for the Crown Prosecution Service. A number of public bodies--the National Crime Agency, the National Economic Crime Centre (NECC), the Cabinet Office--have current interests in fraud research during the time of Covid-19.
This research project has two aims. First, it will map the landscape of how fraudulent conduct is changing during the pandemic. This will be achieved in collaboration with the Crown Prosecution Service (CPS) and a series of roundtable discussions with fraud policy-makers and practitioners. Secondly, by critically appraising semi-structured interviews with leading fraud practitioners, it will provide a normative framework for how emerging dishonest conduct ought to be charged and prosecuted at this time.
The project will lead to two major academic articles on fraud during a pandemic (a theory piece and a governance piece), and a policy paper on fraud governance.
This research project has two aims. First, it will map the landscape of how fraudulent conduct is changing during the pandemic. This will be achieved in collaboration with the Crown Prosecution Service (CPS) and a series of roundtable discussions with fraud policy-makers and practitioners. Secondly, by critically appraising semi-structured interviews with leading fraud practitioners, it will provide a normative framework for how emerging dishonest conduct ought to be charged and prosecuted at this time.
The project will lead to two major academic articles on fraud during a pandemic (a theory piece and a governance piece), and a policy paper on fraud governance.
Organisations
- University of Bristol (Lead Research Organisation)
- Alan Turing Institute (Collaboration)
- Home Office (Collaboration)
- Crown Prosecution Service (CPS) (Collaboration)
- National Crime Agency (Collaboration)
- Home Office (Project Partner)
- National Economic Crime Centre (Project Partner)
- Crown Prosecution Service (Project Partner)
People |
ORCID iD |
| Jennifer Collins (Principal Investigator) |
| Description | Key findings -A detailed study of charges recorded at first hearings in the Magistrates' Court found continued reliance on the common law offence of conspiracy to defraud. The less that new forms of fraud, emerging during the Covid-19 pandemic, are captured and mapped, the greater the pull when they come to the criminal justice system for them to be dealt with using an overly broad offence such as conspiracy to defraud. -Reviews of interviews with fraud professionals and a detailed review of emerging uses, this project has found that AI is rapidly facilitating a new 'fraud-prevention' infrastructure, traversing the public/private domain, without clear limits on its interaction with the criminal justice system and with coercive criminal law measures. -A lowering of the bar for information-sharing must be accompanied by scrutiny of the principles at stake. It is not the case that pressure to increase information-sharing between the private sector, tech industry and law enforcement is an unmitigated good. Increased powers to share reports of suspected fraud with law enforcement must be subject to proportionate threshold access to data. The exclusionary effects for individuals suspected of fraud must be considered. The Bank of England's Artificial Intelligence Public-Private Forum's year-long dialogue on AI innovation between public and private sectors was an important development for this reason. Recommendation: Best practice requires that clarity be provided on how private-public partnerships will use feed information about suspected fraud into the criminal justice system (see Economic Crime Plan:Statement of Progress (July 2019-February 2021) 5). Clear limits must be set on when suspected fraud is reported and why. Key to building a principled response is a fast-track review body which audits de-risking decisions in relation to fraud, as well as considering how information necessary for criminal investigations is handled. A fast-track review body would allow for forward-looking analysis of the incentive and disincentive structures in place in relation to new public-private partnerships specifically. For further details on these recommendations, see J. Collins, 'Why Fraud Prevention is Relevant to Criminal Law Theorists' in A. Simester (ed.), Modern Criminal Law (Forthcoming, Hart 2022) [Copy on file with the author]. -The relevance of AI and ML to fraud detection and prevention is set to grow on an upwards trajectory. Use of AI or ML models to prevent fraud must respect core principles (such as privacy, transparency and explainability) if the use of technological advances in this area is to be principled and robust. Particular challenges for building transparency and explainability should be noted in the fraud-prevention context. For example, too much transparency as to how fraud detection mechanisms work may be exploited by fraudsters, allowing circumvention of fraud detection systems. Recommendation: Advocate for testing of AI and ML systems to detect fraud for accuracy and against bias. The Ada Lovelace Institute has recently developed a proposal for a first Algorithmic Impact Assessment in a different context-healthcare. The aim is to promote good governance around application of algorithms to public sector data, enhancing outcomes for affected groups and to ensure the prevention of bias in process and outcomes. Similar impact assessments ought to be undertaken regarding the use of AI and ML technologies to prevent fraud. Recommendation: Fraud can be detected using several different ML approaches. Some systems use a combination of supervised and unsupervised learning to detect fraud. Literature on ML notes that unsupervised systems may be more difficult to explain. The ability to explain the AI used should be prioritised where unsupervised ML is used to prevent fraud. These systems pose a greater threat to the rule of law, insofar as their outputs can be unforeseen and therefore scrutiny of their outputs is necessary. For further details on these recommendations, see J. Collins, 'Why Fraud Prevention is Relevant to Criminal Law Theorists' in A. Simester (ed.), Modern Criminal Law (Forthcoming, Hart 2022) [Copy on file with the author]. -The common law offence of conspiracy to defraud has withstood robust criticisms of its breadth and its compliance with legal certainty. There is still reliance on this common law offence. This is the case despite the breadth of the general fraud offence found in section 1 of the Fraud Act 2006 and assurances that conspiracy to defraud's use would be reduced since the implementation of the Fraud Act 2006. It is still common to see claims about conspiracy to defraud's 'mopping up' value to prosecutors (most recently in the Ministry of Justice's Post-legislative Assessment of the Fraud Act 2006 (June 2022)). Recommendation: That Committee's review of existing offences to target fraud ought to engage with a detailed literature which makes the point that legal certainty is not respected by the conspiracy to defraud offence. For example, the post-Fraud Act 2006 discussion of conspiracy to defraud which held the offence to be inconsistent with legal certainty, on which see the Joint Parliamentary Committee on Human Rights, Fourteenth Report: Review of International Human Rights Instruments (2006), HL.99, HC 264, para. 2.12; Law Commission (LC 276), Fraud (HMSO, 2002). And the Law Commission Law Commission (LC 276), Fraud (HMSO, 2002) para 1.6; Law Commission, Working Papers of 1974 (No.56) and 1988 (No.104); Law Commission (LC 228), Conspiracy to Defraud (HMSO, 1994). Detailed engagement with Article 7 ECHR and its jurisprudence is needed, given that this right continues to be protected in the proposed Bill of Rights Bill. For detailed argument, see J. Collins, 'Testing Conspiracy to Defraud's Resilience' (2021) 11 Criminal Law Review 902-20. In the absence of conspiracy to defraud being repealed or reformulated to better comply with rule of law requirements, when should the offence be charged? Principled prosecutorial discretion does not resolve rule of law deficiencies, but it is also a key part of the rule of law. The current position in England and Wales is that a decision to use conspiracy to defraud as the offence charged rather than Fraud Act 2006 offences is left to prosecutorial discretion. The Attorney General's guidelines state that, where possible, the general fraud offence found in section 1 of the 2006 Act is to be used by prosecutors. The 'expectation' is that the common law offence will 'be used to a significantly lesser extent' post-Fraud Act 2006. The current guidelines ask prosecutors to consider whether the conduct could be prosecuted under statute, such as under the Fraud Act 2006 or the Criminal Law Act 1977. Statutory provisions should be used 'unless there is a good reason for doing otherwise'.92 The guidance that a lawyer experienced in fraud cases should approve charges other than conspiracy to defraud is welcome, given the complexity of dishonesty offences drafted in inchoate mode.3 Recommendation: It should be written into the guidelines that where an alternative full substantive offence could be charged, it should be. If another offence definition covers conduct also covered by conspiracy to defraud, the presumption should be that the statutory offence is preferred. This is to ensure that, so far as possible, judges are able to adjudicate on the true nature of the wrongful conduct. Conspiracy to defraud is significantly indeterminate in both its conduct and fault requirements. The principled approach is to funnel cases into alternative statutory offences to conspiracy to defraud. This funnelling should not be rebutted by an argument that it would be more straightforward and/or robust evidentially to pursue conspiracy to defraud charges. Recommendation: Prosecutors consider whether conduct can only be prosecuted as conspiracy to defraud (i.e. it is necessary to use the offence). Greater specificity is needed as to these exact circumstances. Currently the prosecutorial guidelines give a non-exhaustive list, including at least one scenario (allowing a bank account to be used by a third party in order to transfer funds) which could now be covered by the Fraud Act 2006 and/or the Serious Crime Act 2015. Examples in the guidance should be withdrawn from the guidance where they are no longer relevant post-implementation of the Fraud Act 2006. Recommendation: The current guidelines state that prosecutors should consider 'whether the available statutory charges adequately reflect the gravity of the offence'. 'Gravity' is currently linked in the guidance to whether the conduct would more effectively be presented as conspiracy to defraud. This permits 'interests of justice' reasoning by more effectively providing an overall picture of fraudulent wrongdoing. To take account of rule-of-law principles, we need clearer guidance on what will furnish strong enough grounds of 'seriousness' for those cases where no other substantive offence exists. Consideration must be given to what harm has been risked and what harm intended as key in determining the seriousness of the conduct. Targeting the vulnerable is another criterion of seriousness. Recommendation: Conspiracy to defraud figures must be published, disaggregated from other charging data. When is the offence charged and what are the figures for successful prosecution? This ought to have been a key recommendation of the review of conspiracy to defraud in 2006. This push for transparency respects rule of law ideals, including Article 7 ECHR's requirements of accessibility and foreseeability. |
| Exploitation Route | Policy recommendations A circumspect approach should be written into prosecutorial guidance on conspiracy to defraud which minimizes conspiracy to defraud's role. This should sit beside a push for greater transparency in the data relating to use of conspiracy to defraud Conspiracy to defraud figures must be published, disaggregated from other charging data. When is the offence charged and what are the figures for successful prosecution? Minutes of the Economic Crime Strategic Board's must be promptly published, and an update provided on the developing overall anti-fraud plan for the UK. Clarity must be provided on how private-public partnerships will use feed information about fraud into the criminal justice system (see Economic Crime Plan: Statement of Progress (July 2019-February 2021) 5). The Artificial Intelligence Public-Private Forum (AIPPF) should consider the particular context of the criminal law, and specifically the offence of fraud, for considering high-risk and high-impact AI models. |
| Sectors | Government, Democracy and Justice,Other |
| Description | Elected as academic member of sub-committee of the Criminal Bar Association given current research interests in fraud, economic crime and emerging tecnologies |
| Geographic Reach | National |
| Policy Influence Type | Participation in a guidance/advisory committee |
| Description | J. Collins, 'Fraud During a Pandemic', Ethics, Law and Governance during Covid-19 (The Pandemic and Beyond, 8 March 2022) |
| Geographic Reach | National |
| Policy Influence Type | Implementation circular/rapid advice/letter to e.g. Ministry of Health |
| Impact | mpact |
| Description | J. Collins, 2022 Sir Gerald Gordon Seminar in Criminal Law (9 June 2022, University of Glasgow) |
| Geographic Reach | Multiple continents/international |
| Policy Influence Type | Influenced training of practitioners or researchers |
| Impact | influence |
| Description | J. Collins, Assize Seminar: Cutting Edge Criminal Law (May 2022, University of Cambridge) |
| Geographic Reach | National |
| Policy Influence Type | Influenced training of practitioners or researchers |
| Impact | impacts |
| Description | Written evidence to the House of Lords' Fraud Act 2006 and Digital Fraud Committee |
| Geographic Reach | Multiple continents/international |
| Policy Influence Type | Contribution to a national consultation/review |
| Impact | Reach and benefits |
| URL | https://committees.parliament.uk/writtenevidence/109591/pdf/ |
| Description | Collaboration with The Alan Turing Institute |
| Organisation | Alan Turing Institute |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Awarded an Alan Turing Institute Fellowship, which began on 1 October 2021. A fellowship with the Turing enables access to a rich and vibrant interdisciplinary data science and AI research community-a key first step in my building a robust interdisciplinary approach to analysing the implication of AI in relation to fraud in the criminal justice system. The fellowship would mark out my long-term intention to partner with innovators in the criminal justice system and business, building effective inter-disciplinary teams which bring fresh ideas to pressing problems. This engagement with multiple stakeholders is key to building adaptable, efficient and principled criminal justice scholarship which has maximum impact in practice. It has become clear that analysis of AI technologies in managing a criminal justice response to the increase in incidence of fraud during the pandemic is necessary |
| Collaborator Contribution | Use of office space in London. Access to inter-disciplinary expertise in AI (ongoing). Attendance at AI UK 2022 conference |
| Impact | Currently active. For example, liaising with other Alan Turing fellows re. the contours of AI and its possible implications in relation to fraud management. |
| Start Year | 2021 |
| Description | Conversations with Home Office |
| Organisation | Home Office |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | Discussion relating to the Home Office's role in leading on COVID response - there has been some protect response from govt but most have been about how citizens were being affected - how to provide advice - and also what effect COVID has had on the fraud demand and how the HO manage the response on fraud. |
| Collaborator Contribution | Sent Conspiracy to Defraud paper Sent round table proposal |
| Impact | Input into conspiracy defraud article; progress on roundtable proposal. |
| Start Year | 2021 |
| Description | Conversations with National Crime Agency |
| Organisation | National Crime Agency |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | Liaising with NEC Fraud Team and specifically vaccine crime during pandemic - also other covid-related fraud. |
| Collaborator Contribution | Academic input on economic crime landscape. |
| Impact | Input into outputs generated for this project. |
| Start Year | 2021 |
| Description | Conversations with the Crown Prosecution Service |
| Organisation | Crown Prosecution Service (CPS) |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | Discussion of Covid and forward-facing economic crime plans. |
| Collaborator Contribution | Academic input on particular challenges to fraud in the economic crime strategy. |
| Impact | Input on output papers for this project. |
| Start Year | 2021 |
| Description | Lead lecture at the 2022 Sir Gerald Gordon Seminar in Criminal Law, University of Glasgow |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | J. Collins, 2022 Sir Gerald Gordon Seminar in Criminal Law (9 June 2022, University of Glasgow) |
| Year(s) Of Engagement Activity | 2022 |
| Description | Pandemic and Beyond: Decision-making During the Pandemic and Beyond |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Pandemic and Beyond hosted, with their policy partners, Culture Commons, relevant stakeholders and decision-makers including those identified by Pandemic and Beyond projects in this cluster to a 'decision-making during the pandemic and beyond panel'. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://pandemicandbeyond.exeter.ac.uk/policy/ |
| Description | Public Lecture at University of Cambridge, involving practitioners, academics and general public |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | J. Collins, Assize Seminar: Cutting Edge Criminal Law (May 2022, University of Cambridge) |
| Year(s) Of Engagement Activity | 2022 |
| Description | Twitter account @FraudPandemicUK |
| Form Of Engagement Activity | Engagement focused website, blog or social media channel |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Active social media account @FraudPandemicUK, allowing for engagement with relevant policymakers and also dissemination of research findings. |
| Year(s) Of Engagement Activity | 2021,2022 |
| URL | https://twitter.com/FraudPandemicUK |