TELUS Chief Data & Trust Officer on her team’s mission to get AI ‘right’

In a recent TM Forum survey of executives from 73 telcos worldwide, 80% identified privacy and security as a big risk for their organization when using GenAI. The same survey also showed that operators are putting in place initiatives to benefit from GenAI, whether it’s establishing centers of excellence, creating use cases in-house, or partnering with technology companies.

So, maybe it should come as no surprise that TELUS’ Chief Data & Trust Officer, Pamela Snively – whose remit includes the company’s privacy program, data ethics, responsible AI, and data governance – has had a very busy 18 months.

“This is potentially an inflection point for humanity and it is moving fast. And that means we need to move quickly as well,” says Snively. “We don’t sleep a lot,” she jokes. But even though the rise of GenAI means her role has “absolutely become more challenging,” it has also had a galvanizing effect on her team and TELUS overall.

“Our team has a very strong sense of a mission,” says Snively. “It’s a very, very exciting space to be in, and it has really highlighted how important it is to get this right.”

TELUS is not fresh to seeking a principled way to deploy AI; it set up its Responsible AI program in 2019. GenAI, however, is changing what Snively’s team – and the rest of the business – can do.

“We’ve been talking for decades about Privacy by Design … and building in controls from the outset. And we have embraced this,” says Snively. “But in many ways, most of the core technology had already been built. ”

Use what you have

Now, however, the newness of GenAI technology combined with an iterative approach to building tools means that TELUS is able to incorporate privacy into GenAI applications from the outset. Even so, Snively recommends pragmatism when considering how to apply privacy and data governance to GenAI.

“We are trying not to reinvent the wheel in terms of governance,” she explains. “We have a very sophisticated privacy and data governance program, we are experienced in impact assessments, and we've been doing that for years in terms of privacy, so we are leveraging that exact same model.”

This approach allows business units to innovate swiftly.

“It’s the same front door for the business. They leverage the existing governance processes, rather than us saying, “Oh, you've got this cool, new AI governance, and we’re going to build something from the ground up,” says Snively. “That takes years. And we don’t have years.”

AI squads

In other instances, organizational changes are accelerating AI experimentation. Snively points out, for example, “that for sales to move as quickly as we want, we needed to break down silos. It couldn’t be led solely out of the analytics team or my team.”

TELUS set up interdisciplinary AI squads that combine business, security, data and analytics expertise to experiment with new Gen AI business cases and tools.

“Setting up a walled garden and safe spaces in which to experiment with this technology was really critical and allowed us to move really quickly in setting up … preliminary technology platforms.”

Other efforts to overcome operational obstacles include building privacy and data protection guardrails around GenAI tools as they are being developed.

“It’s iterative; we might actually put guardrails on that are too strong, and we don't get the utility that we need … and then we figure out what that balance is,” explains Snively.

TELUS’ guardrails also extend to its choice of partners. “We ask: will they help build our customers’ and citizens’ trust in the digital ecosystem? The company also places limits on the data it shares and why, as well as how it’s used. That’s a critical part of what my team does,” Snively adds.

Operationalizing AI

Another focus for Snively is on operationalizing AI responsibly. TELUS is already a member of Responsible AI Institute, but Snively would like to see more cross-sector collaboration between companies, including within the telecoms sector, to identify how AI fits into and improves operations.

“We’ve had a lot of talks about what the challenges are [with AI] and I’m kind of done with that,” says Snively. Instead, she wants to focus on: “How we can fix it? What's working operationally? I’d like to see more of those discussions, and coalescence around some key operational controls that are really working nationally.”

GenAI could even play a role in automating operational controls.

“I'd like to be in a situation where we have more tooling in place that is less labor intensive .”

In the meantime, TELUS is continuing to work on building public trust in AI systems.

“For us, it’s as important to be trusted as to be trustworthy. So, we want to get it right. But we also need to make sure that people know we’re getting it right, because otherwise they won’t be confident participating in the digital ecosystem in AI solutions,” says Snively. “The outcome I want to see is greater customer trust in our programs, and people feeling really comfortable about using AI that TELUS has put out there.”