Achieving a Data-Driven Risk Assessment Methodology for Ethical AI

Abstract

The AI landscape demands a broad set of legal, ethical, and societal considerations to be accounted for in order to develop ethical AI (eAI) solutions which sustain human values and rights. Currently, a variety of guidelines and a handful of niche tools exist to account for and tackle individual challenges. However, it is also well established that many organizations face practical challenges in navigating these considerations from a risk management perspective within AI governance. Therefore, new methodologies are needed to provide a well-vetted and real-world applicable structure and path through the checks and balances needed for ethically assessing and guiding the development of AI. In this paper, we show that a multidisciplinary research approach, spanning cross-sectional viewpoints, is the foundation of a pragmatic definition of ethical and societal risks faced by organizations using AI. Equally important are the findings of cross-structural governance for implementing eAI successfully. Based on evidence acquired from our multidisciplinary research investigation, we propose a novel data-driven risk assessment methodology, entitled DRESS-eAI. In addition, through the evaluation of our methodological implementation, we demonstrate its state-of-the-art relevance as a tool for sustaining human values in the data-driven AI era.

Investigating accountability for Artificial Intelligence through risk governance: A workshop-based exploratory study

“…Effectuation of the defined principles is ensured through risk assessments and management strategies. For example, checklists (e.g., AI HLEG, 2020; Algorithm Watch, 2021b ), impact assessments (e.g., Ada Lovelace Institute, AI Now, and Open Government Partnership, 2021 ); or technical tools (e.g., Vakkuri et al, 2021 ; Felländer et al, 2022 ) are often proposed. They are encouraged to “build on existing policies and governance structures, use pragmatic and action-oriented terminology, focus on risk management in development and procurement, and empower employees through continuous education and change management” ( Mökander et al, 2022 ).…”

Towards Risk-Free Trustworthy Artificial Intelligence: Significance and Requirements

“…Te development of unbiased models must address the defnition of "fairness" [368,369] and determine the appropriate trade-ofs between fairness and performance [369,370]. Furthermore, fairly designed algorithms, hypothetically, may become biased over time when these algorithms are applied in diferent contexts (from what they are designed in) or continuously trained on data with uncorrected biases in a broader healthcare system [359].…”

Descripción de los riesgos y desafíos para la integridad académica de aplicaciones generativas de inteligencia artificial

“…V. RIESGOS La identificación de los riesgos éticos, su evaluación y mitigación son áreas que han adquirido especial relevancia a propósito de la energía que ha mostrado el desarrollo de la IA después de su segundo invierno (Boddington, 2017;Bryson, 2020;Donath, 2020;Eitel-Porter, 2021;Felländer et al, 2022;Hagendorff, 2020;Hermansson & Hansson, 2007;Powers & Ganascia, 2020;Siau & Wang, 2020;Spivak & Shepherd, 2021;Vanem, 2012;Yapo & Weiss, 2018). En general, se destacan los riesgos que plantea la IA para la vigencia de los derechos fundamentales (Council of Europe, 2018;Yeung, 2019), o para la decisión y ejecución de políticas públicas que podrían imponer una algocracia (Danaher, 2016;Chace, 2018, pp.…”