[Paper Review] Accessible Banking: Experiences and Future Directions

Bela Gor, David Aspinall|arXiv (Cornell University)|Jul 16, 2015

Digital Platforms and Economics3 references3 citations

TL;DR

This paper investigates accessibility challenges in digital banking for disabled and older users, identifying critical usability issues in security systems like on-screen keyboards and CAPTCHAs. It proposes inclusive solutions—such as voice biometrics, talking ATMs, sign video banking, and discreet beaconing—demonstrating that redesigning security for accessibility can improve both usability and security without compromising safety.

ABSTRACT

This is a short position paper drawing on experience working with the UK banking industry and their disabled and ageing customers in the Business Disability Forum, a UK non-profit member organisation funded by a large body of UK private and public sector businesses. We describe some commonly reported problems of disabled customers who use modern banking technologies, relating them to UK law and best practice. We describe some of the recent banking industry innovations and the hope they may offer for improved inclusive and accessible multi-channel banking.

Motivation & Objective

• To identify and analyze usability and accessibility barriers in modern banking systems for disabled and older adults.
• To examine how current security mechanisms—such as on-screen keyboards and CAPTCHAs—create exclusion despite regulatory requirements.
• To evaluate emerging technological solutions that aim to improve accessibility without compromising security.
• To explore the role of staff training and organizational culture in maintaining accessible banking services.
• To advocate for embedding accessibility into design from the outset, rather than as an afterthought or reactive fix.

Proposed method

• Conducted qualitative analysis of user experiences from 350 disabled individuals via market research (Missing Out report).
• Collected anecdotal evidence and case studies from Business Disability Forum (BDF) on real-world banking accessibility failures.
• Evaluated emerging technologies such as voice biometrics, talking ATMs, and Bluetooth-based discreet beaconing systems.
• Analyzed existing legal frameworks (e.g., Equality Act 2010) and their implementation gaps in banking services.
• Assessed the impact of software updates and system changes on accessibility, particularly for assistive technology users.
• Proposed design principles for inclusive systems, emphasizing user control, standards compliance, and automated accessibility testing.

Experimental results

Research questions

• RQ1 How do current banking security mechanisms disproportionately affect disabled and older users?
• RQ2 To what extent do well-intentioned security measures like on-screen keyboards and CAPTCHAs undermine accessibility?
• RQ3 What are the real-world impacts of inaccessible banking systems on user trust, financial inclusion, and fraud risk?
• RQ4 How can emerging technologies like voice biometrics and beaconing improve accessibility without increasing security vulnerabilities?
• RQ5 What systemic changes—such as staff training and design-by-default practices—are needed to sustain accessibility in banking services?

Key findings

• Over 80% of disabled and older users face significant barriers in digital banking, particularly with on-screen keyboards and CAPTCHAs.
• Many accessible banking websites become inaccessible after updates due to lack of testing and user control over changes.
• Banks often reject third-party assistance during phone banking, even when users are blind and need help reading screen prompts.
• Voice biometrics and talking ATMs have shown strong positive user feedback, with one blind user describing a talking ATM as 'life-changing'.
• Discreet beaconing systems allow users to communicate accessibility needs to staff without repeated explanation, improving service efficiency and dignity.
• Despite legal obligations under the Equality Act 2010, accessibility is often deprioritized due to lack of staff training and reactive rather than proactive design.