At Dubber, we harness the power of Artificial Intelligence (AI) to analyse conversations, uncovering key insights and pivotal moments to drive your success.
Imagine having a highly efficient partner capable of seamlessly dissecting numerous conversations simultaneously, pinpointing precise information that revolutionises your operations.
That's precisely what our AI achieves.
Our Commitment to Responsible AI Use
Your trust is our top priority, and we've built our AI technology on a foundation of responsible principles. Here's a detailed look at our approach:
Beneficial Purpose: Dubber empowers organisations and their customers with insights that elevate customer experiences and drive business performance.
The Dubber Intelligence Platform minimises environmental footprint by optimising algorithm complexity and dynamically scaling computational resources in response to real-time demands.
Privacy & Security: Dubber maintains strict adherence to ISO/IEC 27001 and GDPR standards to ensure robust data security. We have a designated Data Privacy Officer who independently monitors compliance.
We regularly update the list of Dubber AI sub-processors on our website, reviewing their terms to guarantee data privacy compliance.
Our AI models are built on generalised data and do not incorporate customer-specific information. Customer feedback provided for service enhancement undergoes thorough review to enhance model performance while safeguarding data privacy and adhering to security protocols.
Reliability and Safety: Our AI delivers highly accurate insights through robust quality control mechanisms at every stage of our processes, including:
- Speech-to-text transcription, assessed by Word Error Rate.
- Moment Detection, evaluated using Precision and Recall metrics.
- Moment Topic Classification, measured by Classification Accuracy and Confusion Matrix analysis.
To maintain reliability, aggregated statistics from AI outputs, such as Moment detection rates and topic distributions, undergo continuous monitoring via automated processes. Deviations from expected norms trigger alerts.
While we strive for accuracy, we acknowledge that AI can occasionally make mistakes, so we encourage users to verify outputs against the original conversation.
Fairness: We vigilantly monitor and rectify biases to provide equitable treatment for all users. We evaluate AI modules on balanced datasets, considering factors such as gender, language, regional dialects, and industry. The Dubber Insights product allows users to provide feedback on AI accuracy, with automated monitoring detecting any deviations from the norm.
Users can report perceived bias directly to the Dubber Responsible AI Committee for investigation at trust.engineering@dubber.net.
Transparency: We believe in transparency when it comes to AI-generated information. Our product interface is designed to clearly indicate where AI-generated insights are displayed, enabling users to easily navigate to the original conversation for a deeper understanding of how information was derived.
Accountability & Governance: Our Responsible AI Principles are integral to Dubber’s corporate Environmental, Social, and Governance Policy, accessible through the Dubber Investor Centre.
Through our Information Security and Privacy Governance program, we conduct Data Privacy Impact assessments (DPIAs) as part of our risk-based approach for cyber and data privacy reviews. This includes a thorough review of our data protection controls, technical and security measures, and responsible and ethical use of AI. Operational oversight is carried out by our Responsible AI Committee, accessible to customers at trust.engineering@dubber.net.
For more information about our commitment to responsible AI use and our practices, please visit our Trust Centre.