Regulating Artificial Intelligence

Rademacher, Timo; Wischmeyer, Thomas

doi:10.1007/978-3-030-32361-5

Cited by 40 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This might be done, for example, in cases of investigations and/or audits by regulators. The proposed investigatory and/ or auditing role of regulators in machine learning oversight has been widely advocated for [31,51,76,78]. Although at first logical, the proposed use of post-hoc explainability methods in this capacity suffers from two shortcomings.…”

Section: Post-hoc Explainability Methods and Non-discrimination Law: ...mentioning

confidence: 99%

Explainable artificial intelligence (XAI) post-hoc explainability methods: risks and limitations in non-discrimination law

2022

View full text Add to dashboard Cite

Organizations are increasingly employing complex black-box machine learning models in high-stakes decision-making. A popular approach to addressing the problem of opacity of black-box machine learning models is the use of post-hoc explainability methods. These methods approximate the logic of underlying machine learning models with the aim of explaining their internal workings, so that human examiners can understand them. In turn, it has been alluded that the insights from posthoc explainability methods can be used to help regulate black-box machine learning. This article examines the validity of these claims. By examining whether the insights derived from post-hoc explainability methods in post-model deployment can prima facie meet legal definitions in European (read European Union) non-discrimination law, we argue that machine learning post-hoc explanation methods cannot guarantee the insights they generate. Ultimately, we argue that the use of post-hoc explanatory methods is useful in many cases, but that these methods have limitations that prohibit reliance as the sole mechanism to guarantee fairness of model outcomes in high-stakes decision-making. By way of an ancillary function, the inadequacy of European Non-Discrimination Law for algorithmic decision-making is demonstrated too.

show abstract

Section: Post-hoc Explainability Methods and Non-discrimination Law: ...mentioning

confidence: 99%

Explainable artificial intelligence (XAI) post-hoc explainability methods: risks and limitations in non-discrimination law

2022

View full text Add to dashboard Cite

show abstract

“…48 A sandbox is not a novel concept in the world of computer science, where it denotes "an isolated environment meant for testing and/or preventing malicious programs from damaging a computer system or critical system resources". 49 Its testing function and risk mitigation are key features that bring this concept into the realm of regulation, where it becomes "a process and a tool for regulation", comparable to a laboratory where innovations are tested against the existing regulatory framework, through a process involving participating business entities and a regulator. 50 Together with innovation hubs, 51 There is no uniform definition of innovation hubs, but they typically encompass onestop shops involving multi-stakeholder cooperation and serve as a "…central contact point to streamline queries and provide support, advice, and guidance to either regulated or unregulated firms, helping them navigate the regulatory, supervisory, policy, or legal environment.…”

Section: The Concept and Features Of Regulatory Sandboxesmentioning

confidence: 99%

Regulatory Sandboxes Under the Draft Eu Artificial Intelligence Act: An Opportunity for Smes?

Pošćić¹,

Martinović

2023

iele

View full text Add to dashboard Cite

More than a year after the European Commission submitted the Proposal for the Artificial Intelligence Act (AI Act), the EU institutions are still working on adopting this groundbreaking regulation. The Draft AI Act contains a uniform set of horizontal rules for the development, marketing, and use of AI systems in conformity with the Union values, applying the proportionate risk-based approach. The aim is to avoid regulatory friction and fragmentation and to create a well-functioning internal market for AI systems and technologies. However, the policy and regulatory choices should not obstruct the innovative potential and transformative impact of AI systems and technologies on the society and economy. The Draft AI Act, therefore, introduces AI regulatory sandboxes, as a testing ground for deciding what to regulate and how. This is a novel regulatory approach, fostering innovation, development, and testing of AI systems under strict regulatory oversight before these systems are placed on the market. The proposed solution from the Draft AI Act has caused both excitement and criticism in the legal doctrine and industry. This paper will explore the benefits and challenges of AI regulatory sandboxes. The draft provisions will be critically evaluated, drawing from the experience in the FinTech industry, especially considering the effect on SMEs. The EU's ambition is to set up a robust and disruption-resilient, yet flexible, innovation-friendly, and future-proof regulatory framework for AI, and the intuitive appeal of AI regulatory sandboxes for both regulators and innovators deserves an in-depth examination.

show abstract

“…151 Therefore, it is foremost to consider AI's transparency as balancing interests and governance issues, which require adequate attention to multidisciplinary development. 152 However, it should be marked that there is not always a "onesize-fits-all" approach, followingly, not everyone needs free access to every piece of information 153 , or there is no need to disclose all the information either since they may consider trade secret. Besides, sometimes, for various reasons, information cannot be disclosed, for example, if training data contains "Black Box" algorithms, as mentioned in the different part of this Thesis.…”

Section: Is Gdpr Adequate In Ensuring Privacy and Transparency In An ...mentioning

confidence: 99%

Privacy and Transparency in an AI-driven World: Does Algorithmic Transparency Fit on Data Privacy Under GDPR?

Gasimova

2023

SSRN Journal

View full text Add to dashboard Cite

Regulating Artificial Intelligence

Cited by 40 publications

References 0 publications

Explainable artificial intelligence (XAI) post-hoc explainability methods: risks and limitations in non-discrimination law

Explainable artificial intelligence (XAI) post-hoc explainability methods: risks and limitations in non-discrimination law

Regulatory Sandboxes Under the Draft Eu Artificial Intelligence Act: An Opportunity for Smes?

Privacy and Transparency in an AI-driven World: Does Algorithmic Transparency Fit on Data Privacy Under GDPR?

Contact Info

Product

Resources

About