Library

@article{koshiyama_towards_2024,

title = {Towards algorithm auditing: managing legal, ethical and technological risks of AI, ML and associated algorithms},

author = {Adriano Koshiyama and Emre Kazim and Philip Treleaven and Pete Rai and Lukasz Szpruch and Giles Pavey and Ghazi Ahamat and Franziska Leutner and Randy Goebel and Andrew Knight and Janet Adams and Christina Hitrova and Jeremy Barnett and Parashkev Nachev and David Barber and Tomas Chamorro-Premuzic and Konstantin Klemmer and Miro Gregorovic and Shakeel Khan and Elizabeth Lomas and Airlie Hilliard and Siddhant Chatterjee},

url = {https://royalsocietypublishing.org/doi/10.1098/rsos.230859},

doi = {10.1098/rsos.230859},

year  = {2024},

date = {2024-05-01},

urldate = {2024-10-21},

journal = {Royal Society Open Science},

volume = {11},

number = {5},

pages = {230859},

abstract = {­Business reliance on algorithms is becoming ubiquitous, and companies are increasingly concerned about their algorithms causing major financial or reputational damage. High-profile cases include Google’s AI algorithm for photo classification mistakenly labelling a black couple as gorillas in 2015 (Gebru 2020 In The Oxford handbook of ethics of AI, pp. 251–269), Microsoft’s AI chatbot Tay that spread racist, sexist and antisemitic speech on Twitter (now X) (Wolf et al. 2017 ACM Sigcas Comput. Soc. 47, 54–64 (doi:10.1145/3144592.3144598)), and Amazon’s AI recruiting tool being scrapped after showing bias against women. In response, governments are legislating and imposing bans, regulators fining companies and the judiciary discussing potentially making algorithms artificial ‘persons’ in law. As with financial audits, governments, business and society will require algorithm audits; formal assurance that algorithms are legal, ethical and safe. A new industry is envisaged: Auditing and Assurance of Algorithms (cf. data privacy), with the remit to professionalize and industrialize AI, ML and associated algorithms. The stakeholders range from those working on policy/regulation to industry practitioners and developers. We also anticipate the nature and scope of the auditing levels and framework presented will inform those interested in systems of governance and compliance with regulation/standards. Our goal in this article is to survey the key areas necessary to perform auditing and assurance and instigate the debate in this novel area of research and practice.},

note = {Publisher: Royal Society},

keywords = {AI and Financial Systems},

pubstate = {published},

tppubtype = {article}

}

@article{shivram_auditing_2024,

title = {Auditing with Ai: A Theoretical Framework for Applying Machine Learning Across the Internal Audit Lifecycle},

author = {Vivek Shivram},

url = {https://doi.org/10.1080/07366981.2024.2312025},

doi = {10.1080/07366981.2024.2312025},

issn = {0736-6981},

year  = {2024},

date = {2024-01-01},

urldate = {2024-10-21},

journal = {EDPACS},

volume = {69},

number = {1},

pages = {22–40},

abstract = {Artificial Intelligence (hereinafter AI), and specifically, Machine Learning (hereinafter, ML), has shown tremendous potential to revolutionize the internal audit (hereinafter, IA) profession, from enabling audit coverage of entire test populations, to introducing objectivity in the analysis of key areas. However, prior literature shows that the multiplicity of innovation options can be overwhelming. This paper aims to offer a theoretical framework that would enable audit practitioners, within both industry and professional services, to consider how ML capabilities can be harnessed to their fullest potential across the internal audit lifecycle, from audit planning to reporting. The paper discusses how DA and ML capabilities relate to the internal audit function’s (hereinafter, IAF) remit, drawing from extant literature. In doing so, the paper identifies the most specific options available to IAFs to drive innovation across each segment of the audit lifecycle, leveraging various DA and ML techniques, and supports the assertion that auditors require a continuous innovation mind-set to be effective change agents. The paper also draws out the requirement for effective guardrails, especially with emerging technology, such as Generative AI. Finally, the paper discusses how the value arising from these efforts can be measured by IAFs.},

note = {Publisher: Taylor & Francis 

_eprint: https://doi.org/10.1080/07366981.2024.2312025},

keywords = {AI and Financial Systems},

pubstate = {published},

tppubtype = {article}

}

@article{commerford_man_2022,

title = {Man Versus Machine: Complex Estimates and Auditor Reliance on Artificial Intelligence},

author = {Benjamin P. Commerford and Sean A. Dennis and Jennifer R. Joe and Jenny W. Ulla},

url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/1475-679X.12407},

doi = {10.1111/1475-679X.12407},

issn = {1475-679X},

year  = {2022},

date = {2022-01-01},

urldate = {2024-10-21},

journal = {Journal of Accounting Research},

volume = {60},

number = {1},

pages = {171–201},

abstract = {Audit firms are investing billions of dollars to develop artificial intelligence (AI) systems that will help auditors execute challenging tasks (e.g., evaluating complex estimates). Although firms assume AI will enhance audit quality, a growing body of research documents that individuals often exhibit “algorithm aversion”—the tendency to discount computer-based advice more heavily than human advice, although the advice is identical otherwise. Therefore, we conduct an experiment to examine how algorithm aversion manifests in auditor judgments. Consistent with theory, we find that auditors receiving contradictory evidence from their firm's AI system (instead of a human specialist) propose smaller adjustments to management's complex estimates, particularly when management develops their estimates using relatively objective (vs. subjective) inputs. Our findings suggest auditor susceptibility to algorithm aversion could prove costly for the profession and financial statements users.},

note = {_eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/1475-679X.12407},

keywords = {AI and Financial Systems},

pubstate = {published},

tppubtype = {article}

}

@article{azzutti_machine_2021,

title = {Machine Learning, Market Manipulation, and Collusion on Capital Markets: Why the "Black Box" Matters},

author = {Alessio Azzutti and Wolf-Georg Ringe and H. Siegfried Stiehl},

url = {https://heinonline.org/HOL/P?h=hein.journals/upjiel43&i=83},

year  = {2021},

date = {2021-01-01},

urldate = {2024-10-21},

journal = {University of Pennsylvania Journal of International Law},

volume = {43},

number = {1},

pages = {79–136},

abstract = {This Article offers a novel perspective on the implications of increasingly autonomous and “black box” algorithms, within the ramification of algorithmic trading, for the integrity of capital markets. Artificial intelligence (AI) and particularly its subfield of machine learning (ML) methods have gained immense popularity among the great public and achieved tremendous success in many real-life applications by leading to vast efficiency gains. In the financial trading domain, ML can augment human capabilities in price prediction, dynamic portfolio optimization, and other financial decision-making tasks. However, thanks to constant progress in the ML technology, the prospect of increasingly capable and autonomous agents to delegate operational tasks and even decision-making is now beyond mere imagination, thus opening up the possibility for approximating (truly) autonomous trading agents anytime soon. 

Given these spectacular developments, this Article argues that such autonomous algorithmic traders may involve significant risks to market integrity, independent from their human experts, thanks to self-learning capabilities offered by state-of-the-art and innovative ML methods. Using the proprietary trading industry as a case study, we explore emerging threats to the application of established market abuse laws in the event of algorithmic market abuse, by taking an interdisciplinary stance between financial regulation, law and economics, and computational finance. Specifically, our analysis focuses on two emerging market abuse risks by autonomous algorithms: market manipulation and “tacit” collusion. We explore their likelihood to arise in global capital markets and evaluate related social harm as forms of market failures. 

With these new risks in mind, this Article questions the adequacy of existing regulatory frameworks and enforcement mechanisms, as well as current legal rules on the governance of algorithmic trading, to cope with increasingly autonomous and ubiquitous algorithmic trading systems. We demonstrate how the “black box” nature of specific ML-powered algorithmic trading strategies can subvert existing market abuse laws, which are based upon traditional liability concepts and tests (such as “intent” and “causation”). We conclude by addressing the shortcomings of the present legal framework and develop a number of guiding principles to assist legal and policy reform in the spirit of promoting and safeguarding market integrity and safety.},

keywords = {AI and Financial Systems},

pubstate = {published},

tppubtype = {article}

}

@article{lu_algorithmic_2020,

title = {Algorithmic Opacity, Private Accountability, and Corporate Social Disclosure in the Age of Artificial Intelligence},

author = {Sylvia Lu},

url = {https://heinonline.org/HOL/P?h=hein.journals/vanep23&i=111},

year  = {2020},

date = {2020-01-01},

urldate = {2024-10-21},

journal = {Vanderbilt Journal of Entertainment & Technology Law},

volume = {23},

number = {1},

pages = {99–160},

abstract = {Today, firms develop machine learning algorithms in nearly every industry to control human decisions, creating a structural tension between commercial opacity and democratic transparency. In many forms of business applications, advanced algorithms are technically complicated and privately owned, hiding from legal regimes and preventing public scrutiny, although they may demonstrate their erosion of democratic norms, damages to financial gains, and extending harms to stakeholders without warning. Nevertheless, because the inner workings and applications of algorithms are generally incomprehensible and protected as trade secrets, they can be completely shielded from public surveillance. One of the solutions to this conflict between algorithmic opacity and democratic transparency is an effective mechanism that incentivizes firms to engage in information disclosure for their algorithms. 

This Article argues that the pressing problem of algorithmic opacity is due to the regulatory void of US disclosure regulations that fail to consider the informational needs of stakeholders in the age of AI. In a world of privately-owned algorithms, advanced algorithms as the primary source of decision-making power have produced various perils for the public and firms themselves, particularly in the context of the capital market. While the current disclosure framework has not considered the informational needs associated with algorithmic opacity, this Article argues that algorithmic disclosure under securities law could be used to promote private accountability and further public interest in sustainability. 

First, as I discuss, advanced machine learning algorithms have been widely applied in AI systems in many critical industries, including financial services, medical services, and transportation services. Second, despite the growing pervasiveness of algorithms, the laws, particularly intellectual property laws, continue to encourage the existence of algorithmic opacity. Although the protection of trade secrecy in algorithms seems beneficial for firms to create competitive advantage, as I examine, it has proven deleterious for society, where democratic norms such as privacy, equality, and safety are now being compromised by invisible algorithms that no one can ever scrutinize. Third, although the emerging perils of algorithmic opacity are much more catastrophic and messier than before, the current disclosure framework in the context of corporate securities laws fails to consider the informational needs of the stakeholders for advanced algorithms in AI systems. 

In this vein, through the lens of the US Securities and Exchange Commission (SEC) disclosure framework, this Article proposes a new disclosure framework for machine-learning-algorithm-based AI systems that considers the technical traits of advanced algorithms, potential dangers of AI systems, and regulatory governance systems in light of increasing AI incidents. Towards this goal, I discuss numerous disclosure topics, analyze key disclosure reports, and propose new principles to help reduce algorithmic opacity, including stakeholder consideration, sustainability consideration, comprehensible disclosure, and minimum necessary disclosure, which I argue can ultimately strike a balance between democratic values in transparency and private interests in opacity. This Article concludes with a discussion of the impacts, limitations, and possibilities of using the new disclosure framework to promote private accountability and corporate social responsibility in the AI era.},

keywords = {AI and Financial Systems},

pubstate = {published},

tppubtype = {article}

}

@article{johnson_artificial_2019,

title = {Artificial Intelligence, Machine Learning, and Bias in Finance: Toward Responsible Innovation Symposium: Rise of the Machines: Artificial Intelligence, Robotics, and the Reprogramming of Law},

author = {Kristin Johnson and Frank Pasquale and Jennifer Chapman},

url = {https://heinonline.org/HOL/P?h=hein.journals/flr88&i=515},

year  = {2019},

date = {2019-01-01},

urldate = {2024-10-21},

journal = {Fordham Law Review},

volume = {88},

number = {2},

pages = {499–530},

abstract = {Over the last decade, a growing number of digital startups launched bids to lure business from the financial services industry. Financial technology (“fintech”) firms deploying ever more complex and opaque algorithms assess the creditworthiness of consumers. Armed with vast quantities of data and complex algorithms to interpret the data, these firms are reigniting debates about how best to regulate financial institutions and technology firms engaged in consumer banking activities. 

With a few quick taps on a smart phone, consumers can access a growing universe of apps that offer discounted interest rates on consumer loans. For proponents, the launch of fintech firms marks a new frontier in the ever-expanding utopian vision of the “technological sublime” or faith-like devotion to the potential for technology to transform us into a more equitable and just society. Consumer advocates are justifiably skeptical. While legally prohibited today, well-documented discriminatory, exclusionary, and predatory credit market practices persist. 

This Essay describes fintech firms’ integration of learning algorithms and their anticipated economic and social welfare benefits — enhanced efficiency, accuracy, and accessibility. We then examine the emerging regulatory landscape. Over the last decade, federal banking regulators signaled and adopted policies that preempted state regulatory authority over fintech firms. A recent announcement by the Office of the Comptroller of the Currency (OCC) revealed the agency’s intention to allow fintech firms to apply for special purpose charters that would permit them to operate, in many respects, as national banks (“Fintech Charter Decision”). 

The OCC’s Fintech Charter Decision creates gaps in the supervision of fintech firms and encourages market participants to engage in regulatory arbitrage. We argue that federal special purpose charters set the stage for regulatory arbitrage and may enable fintech firms to minimize their exposure to state antidiscrimination and consumer protection regulations. Reducing regulatory oversight of these important legal and ethical norms in a dynamic and evolving market defined by a technology that may import unconscious biases and disadvantage lower-income individuals and families raises red flags. We conclude with brief reflections regarding the necessity for courts and regulators to balance the promised benefits of fintech firms’ neo-banking initiatives with the historic and special gatekeeping role of banking platforms. Unilateral deregulatory action by state or federal regulators may undermine efforts to ensure effective oversight of fintech firms that seek to extend access to safe and affordable banking services.},

keywords = {AI and Financial Systems},

pubstate = {published},

tppubtype = {article}

}

@article{lin_artificial_2019,

title = {Artificial Intelligence, Finance, and the Law Symposium: Rise of the Machines: Artificial Intelligence, Robotics, and the Reprogramming of Law},

author = {Tom C. W. Lin},

url = {https://heinonline.org/HOL/P?h=hein.journals/flr88&i=547},

year  = {2019},

date = {2019-01-01},

urldate = {2024-10-21},

journal = {Fordham Law Review},

volume = {88},

number = {2},

pages = {531–552},

abstract = {Artificial intelligence is an existential component of modern finance. The progress and promise realized and presented by artificial intelligence in finance has been thus far remarkable. It has made finance cheaper, faster, larger, more accessible, more profitable, and more efficient in many ways. Yet for all the significant progress and promise made possible by financial artificial intelligence, it also presents serious risks and limitations. This Article offers a study of those risks and limitations—the ways artificial intelligence and misunderstandings of it can harm and hinder law, finance, and society. It provides a broad examination of inherent and structural risks and limitations present in financial artificial intelligence, explains the implications posed by such dangers, and offers some recommendations for the road ahead. Specifically, it highlights the perils and pitfalls of artificial codes, data bias, virtual threats, and systemic risks relating to financial artificial intelligence. It also raises larger issues about the implications of financial artificial intelligence on financial cybersecurity, competition, and society in the near future. Ultimately, this Article aspires to share an insightful perspective for thinking anew about the wide-ranging effects at the intersection of artificial intelligence, finance, and the law with the hopes of creating better financial artificial intelligence—one that is less artificial, more intelligent, and ultimately more humane, and more human.},

keywords = {AI and Financial Systems},

pubstate = {published},

tppubtype = {article}

}