Dive In Festival opens its final edition tomorrow with the insurance industry facing a question that has become increasingly difficult to separate from conversations about talent and inclusion: who benefits as artificial intelligence becomes embedded in everyday work?
The global festival for culture and talent runs from September 22 to 24, marking its 12th and final year. More than 30 countries are expected to participate, with this year’s theme, “The Human gAIn: Powering Culture & Connection,” focused on how insurance organizations can adopt AI without losing the judgment, creativity, trust and connection that remain central to the industry.
It is a timely closing theme for an event that began in London in 2015 and has since developed into a global insurance initiative. Last year alone, Dive In attracted more than 35,000 attendees across 31 countries and 102 events.
For insurance leaders entering this week’s discussions, the debate has already shifted beyond whether AI will become part of the workplace, according to Clare Barley (pictured), managing director at Asta.
“We’re done with AI being a future world,” said Barley. “We’ve moved beyond a world where, particularly from a leadership perspective, you’re just worrying about the technology itself: Can I implement it well? Can I get the right things out of it? It is going to amplify what we do all day, every day.”
Insurers, brokers and other insurance businesses are now confronting questions around how work is distributed, which employees receive access to new tools and training, and whether AI ultimately broadens opportunity or reinforces existing inequalities.
Barley expects trust to feature heavily in the conversations beginning this week, alongside adoption, judgment, guardrails and the skills employees will need as AI becomes more capable.
Among these, she flagged bias as among the biggest issues occupying insurance leaders as deployment accelerates. Insurance organizations working historical data can carry forward patterns, hastened by AI, that may not reflect future conditions or outcomes.
“The past doesn’t necessarily represent the future, so to the extent that you’ve got bias in your data and amplify it forward, I don’t think that’s a new thought, particularly inside the insurance industry,” Barley noted. “It’s what you don’t know is being amplified, and what you don’t know in those biases that’s being brought forward,” she said.
That could surface in ways that have little to do with a visibly discriminatory algorithm. Barley pointed to differences in how quickly teams receive access to new generations of AI, the quality of training they receive and how much opportunity employees have to use the technology in their day-to-day roles.
“One team gets the first, second and third goes at the latest generations of AI while other teams get later adoption, or it could be age-related,” she said. “I don’t know what the demographic split will be, but I’m fairly certain it will be one that’s unexpected.”
Barley is skeptical that organizations will ever eliminate bias completely. “I think bias is inherent. Bias is part of human life. To a degree, there is no segment of anything humans have ever touched that hasn’t created bias at some point,” she said. “To think we can stamp out bias in its entirety is unlikely.”
Governance and guardrails can reduce the risk, but leaders should expect new problems to emerge as existing ones are addressed: “It will be a bit of a whack-a-mole,” Barley continued. “You address one and an unexpected one pops back out somewhere else.”
Monitoring workplace outcomes is as important as reviewing the technology itself. From there, leaders need to trace those differences back to their source, whether that is the technology, training, access or another part of implementation.
And the same technology also has the potential to remove some barriers. Barley said AI can make knowledge easier to access, including information that previously depended heavily on experience, established networks or understanding unwritten workplace conventions. Whether that potential translates into more equitable workplaces may depend on how evenly the technology itself is distributed.
“The beautiful thing about AI is the leveling of the playing field. If you don’t know your stuff, you can find out pretty quickly,” she said. “If you can start from a very clear position of: this is where the future is going, this is where our organization is going, and this is where our people and technology will fit together, you set the ground level for everyone."
