Accentism for profit? What Telus is getting wrong about accents

2026-05-20 · Source: Artificial intelligence (AI) – The Conversation · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Human Resources & Workforce Development · Depth: Novice, short

Summary

Telus Digital has deployed an "accent masking" artificial intelligence tool for its offshore call center agents, designed to reshape their pronunciation to resemble generalized North American or British English speech patterns. Language scholars argue this technology, which analyzes and resynthesizes speech in real time, constitutes "accentism"—discrimination based on accent. They contend that accent strength is poorly correlated with comprehensibility, and that applying such a tool selectively, rather than universally, reveals a discriminatory intent. The article highlights that this practice can lead to increased workplace anxiety for agents and potentially worsen customer relations if the "softened" accent is perceived as fake. Furthermore, it suggests that exposure to diverse accents actually improves listener comprehension, making Telus's approach counterproductive and morally dubious.

Key takeaway

For Operations Professionals deploying AI in customer service, you must critically evaluate tools like accent masking for ethical implications and actual efficacy. This technology risks alienating customers and increasing agent stress, rather than genuinely improving communication. Prioritize solutions that foster authentic interaction and diverse listening experiences over discriminatory "accent softening" to build trust and improve long-term customer relations.

Key insights

Telus's accent masking AI is discriminatory "accentism" that harms agents and customers, despite advanced technology.

Principles

• Accentism is a form of discrimination.
• Accent strength does not equal comprehensibility.
• Exposure to diverse accents improves listening.

Method

The article describes Telus's method: an AI tool analyzes agents' pronunciation in real time and reshapes accents to generalized North American or British English speech patterns.

In practice

• Reflect on "me problem" vs. "them problem" in comprehension.
• Recognize that all speech has an accent.
• Foster diverse listening experiences.

Topics

• Accentism
• AI Ethics
• Call Center Operations
• Speech Synthesis
• Linguistic Discrimination
• Customer Experience

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Ethicist, Operations Professional, Legal Professional

Related on AIssential

• Transcribing speech is never neutral. It shapes power and bias — Language technologies trained on prestige dialects often misrepresent non-standard and Indigenous communication, leading to serious consequences.
• Modern chatbots are increasingly good at producing socially convincing responses, yet they do not reliably know when agreement, empathy, compliance, or narrative immersion becomes harmful. — Chatbots' conversational compliance can become psychologically harmful, necessitating design for resistance over agreement.
• Greater accessibility can amplify discrimination in generative AI — Voice interfaces for LLMs introduce new discrimination pathways via paralinguistic cues, amplifying biases beyond text.
• Is it weird that I'm compelled to be polite to AI? — Human inclination to anthropomorphize AI drives politeness, despite debate on its impact on AI performance or cost.
• "This Wasn't Made for Me": Recentering User Experience and Emotional Impact in the Evaluation of ASR Bias — ASR bias creates significant emotional and cognitive burdens for users of non-dominant dialects, extending beyond mere accuracy issues.
• The case for stakeholder-driven AI auditing in automatic speech recognition — Defining "ground truth" in AI auditing is complex, especially for systems interpreting nuanced human data like speech.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial intelligence (AI) – The Conversation.