Telus Deploys Real-Time Voice Conversion to Alter Call-Centre Agent Accents, Sparks Labour and Privacy Backlash

Summary

Canadian telecommunications company Telus has deployed a real-time speech-to-speech system developed by Tomato.ai through its Telus Digital unit to modify the accents of offshore call-centre agents during live customer calls. The voice conversion technology applies automatic speech recognition, accent conversion models, and neural vocoders at low latency to reduce what Telus describes as "accent-related friction" between agents and customers. The system alters agent voices without explicit disclosure to callers, raising significant concerns about transparency and consent.

The deployment has provoked swift backlash in Canada, with labour groups and privacy advocates criticizing the practice as deceptive and calling for mandatory disclosure requirements. Critics argue the technology masks worker identity and raises questions about consent, worker rights, and voice-privacy regulations in customer-facing services. The controversy highlights broader tensions around real-time voice alteration in contact centres, where operational efficiency must be balanced against ethical deployment and regulatory compliance.

Competitors Rogers and Bell have publicly stated they have no plans to adopt similar voice-altering technology, signalling how public backlash and regulatory scrutiny may constrain broader adoption of voice conversion in Canada's telecommunications sector. The incident reflects growing scrutiny on transparency and worker rights in AI-powered customer service systems.

• Competitors Rogers and Bell have rejected similar technology, suggesting public backlash may limit sector-wide adoption