Not All Datasets Are Created Equal: New TELUS Digital Survey Shows Trust in AI is Dependent on How Data is Sourced

As generative AI (GenAI) continues to accelerate across industries, trust in what type of data is used to train, evaluate and fine-tune AI models is emerging as a critical issue. A new TELUS Digital survey of 1,000 U.S. adults found that nearly 9 in 10 (87%) respondents (up from 75% in 2023) believe companies should be transparent about how they source data for GenAI models. Additionally, 65% believe that the exclusion of high-quality, verified content, such as information from trusted media sources (e.g. New York Times, Reuters, Bloomberg), can lead to inaccurate and/or biased large language model (LLM) responses.

This press release features multimedia. View the full release here: https://www.businesswire.com/news/home/20250624647463/en/

“As AI systems become more specialized and embedded in high-stakes use cases, the quality of the datasets used to optimize outputs is emerging as a key differentiator for enterprises between average performance and having the potential to drive real-world impacts,” said Amith Nair, Global VP and General Manager, Data & AI Solutions, TELUS Digital. “We’re well past the era where general crowdsourced or internet data can meet today’s enterprises’ more complex and specialized use cases. This is reflected in the shift in our clients’ requests from ‘wisdom of the crowd’ datasets to ‘wisdom of the experts’. Experts and industry professionals are helping curate such datasets to ensure they are technically sound, contextually relevant and responsibly built. In high-stakes domains like healthcare or finance, even a single mislabelled data point can distort model behavior in ways that are difficult to detect and costly to correct. TELUS Digital works with a diverse network of experts to ensure datasets reflect a range of perspectives, reduce bias and align more closely with real-world use.”

In response to evolving industry dynamics, TELUS Digital Experience (TELUS Digital) (NYSE and TSX:TIXT), a leading global technology company specializing in digital customer experiences, has launched 13 off-the-shelf STEM (science, technology, engineering and mathematics) datasets, including coding and reasoning data that is critical for LLM advancements. The datasets have been expertly-curated by a diverse pool of contributors, including Ph.D. researchers, professors, graduate students and working professionals from around the world. This gives enterprises access to high-quality data that has been cleaned, labeled and formatted for immediate integration into AI training workflows.

Why does human expertise matter?

In complex fields like STEM, trained experts bring deep contextual understanding of the subject matter to be able to more accurately interpret ambiguous inputs, apply consistent standards, and recognize subtle distinctions such as legal implications or scientific classifications. Experts are also better equipped to identify edge cases and help mitigate the cognitive biases that can undermine model performance.

Dancan, an AI scientist with a background in organic chemistry who is also a freelance data annotator for TELUS Digital, said, “By annotating data properly, we’re enabling AI to better collaborate with scientists, helping them streamline their processes and find solutions faster and at a fraction of the cost. Combining my background in chemistry with the potential of AI helps speed up the discovery of life-saving therapies, which is what I’m so passionate about.”

Sourabh, a software engineer from India and freelance data annotator for TELUS Digital, said, "Coming from a software background, I’ve always been drawn to the problem-solving side of AI. I've been able to apply that mentality to practical annotation projects such as solving coding challenges, where we provide step-by-step explanations that influence how models learn and function."

Justin, a Ph.D. candidate in Chemistry at the University of Vermont, adds, “I have found that high-quality data annotation in conjunction with a well-trained LLM can significantly lower the barrier to entry (or re-entry) for even the strongest, most well-qualified scientist to work on a new project. This is where the data annotation work that I perform with TELUS Digital can truly lead to more efficient and profound scientific innovation.”

More information:

With 20+ years of experience in data annotation, TELUS Digital has built a deep bench of global expertise across complex, high-context domains. In addition to its off-the-shelf datasets, the company delivers fully custom AI data projects tailored to clients’ specific use cases, industry requirements, and linguistic needs. This includes multilingual and multimodal annotation and specialized support for sectors where accuracy and regulatory compliance are essential to model performance.

TELUS Digital’s data annotation services are powered by Ground Truth (GT) Studio, its advanced, proprietary data labeling platform that brings together diverse, global human expertise with intelligent automation to ensure exceptional data quality, accuracy and efficiency. As part of TELUS Digital’s end-to-end data platform:

• Experts Engine connects vetted, domain-specific professionals to annotation and validation tasks, assembling tailored teams based on industry, language and project complexity.
• Fine-Tune Studio automates checks for accuracy, solvability, LaTeX formatting (used for math and scientific notation) and grammar, ensuring every dataset meets high technical and ethical standards.

TELUS Digital’s off-the-shelf STEM datasets can be licensed individually or as a collection, and collectively include over 178,000 structured prompt-response pairs. For projects that demand deeper customization or use-case specificity, the company also delivers bespoke datasets aligned to specific use cases and regulatory requirements. To learn more about TELUS Digital’s off-the-shelf datasets and how they support scalable, real-world AI applications, visit our Data & AI Solutions page.

Survey Methodology: The survey findings are based on a Pollfish survey that was conducted on June 9, 2025, and included responses from 1,000 men and women aged 18+ who live in the United States and indicated a familiarity with generative AI tools.

About TELUS Digital

TELUS Digital (NYSE & TSX: TIXT) crafts unique and enduring experiences for customers and employees, and creates future-focused digital transformations that deliver value for our clients. We are the brand behind the brands. Our global team members are both passionate ambassadors of our clients’ products and services, and technology experts resolute in our pursuit to elevate their end customer journeys, solve business challenges, mitigate risks, and drive continuous innovation. Our portfolio of end-to-end, integrated capabilities include customer experience management, digital solutions, such as cloud solutions, AI-fueled automation, front-end digital design and consulting services, AI & data solutions, including computer vision, and trust, safety and security services. Fuel iX™ is TELUS Digital’s proprietary platform and suite of products for clients to manage, monitor, and maintain generative AI across the enterprise, offering both standardized AI capabilities and custom application development tools for creating tailored enterprise solutions.

Powered by purpose, TELUS Digital leverages technology, human ingenuity and compassion to serve customers and create inclusive, thriving communities in the regions where we operate around the world. Guided by our Humanity-in-the-Loop principles, we take a responsible approach to the transformational technologies we develop and deploy by proactively considering and addressing the broader impacts of our work. Learn more at: telusdigital.com

View source version on businesswire.com: https://www.businesswire.com/news/home/20250624647463/en/