AI Collaboration to Counteract Flaws in High-Stakes Financial Decisions

Sachan, Swati ORCID: 0000-0003-0136-0553 (2025) AI Collaboration to Counteract Flaws in High-Stakes Financial Decisions [Report]

Access the full-text of this item by clicking on the Open Access link.

PDF Written_Evidence_committees.parliament.uk.pdf - Other Download (165kB) | Preview

Executive Summary

Summary of Evidence: Q1: AI-driven financial systems face challenges on decision opacity, sparse or unreliable data, and security vulnerabilities. Additionally, the demands for future scalability require a robust computational infrastructure for equitable outcomes in high-stakes financial domains. Q2: AI outperforms human experts in repetitive (rule-based) tasks but falters when faced with ambiguous data and novel scenarios, whereas financial experts excel with contextual reasoning and ethical judgment. Q3: AI's "bias" is not intrinsic but a reflection of flawed human inputs. Human cognitive limitations: redundant data, biases (predictable deviations), and judgment "noise" (unexplained variability) contaminate AI systems by embedding inconsistent human decisions into training data. Q4: The potential of Decentralised Finance (DeFi), powered by blockchain's immutability (tamper-proof data permanence) and transparency (publicly auditable), to address the scarcity of reliable financial data and accountability in AI-driven systems.

Item Type:	Report
Divisions:	Faculty of Humanities & Social Sciences Faculty of Humanities & Social Sciences > School of Management
Depositing User:	Symplectic Admin
Date Deposited:	21 May 2025 08:21
Last Modified:	21 May 2025 08:21
Open Access URL:	https://committees.parliament.uk/writtenevidence/1...
Related Websites:	Publisher
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3192837
Disclaimer:	The University of Liverpool is not responsible for content contained on other websites from links within repository metadata. Please contact us if you notice anything that appears incorrect or inappropriate.