Insurers using old strategies to tackle ‘futuristic problem’ of fraud

Insurers are attempting to tackle increasingly sophisticated fraud involving deepfakes and synthetic identities using outdated investigative techniques, delegates at the Insurance Fraud Investigators Group (IFIG) conference heard.

Speaking during a conference panel titled The overlooked channel: Why telephony deserves a place in your fraud investigation toolkit, Adele Sumner, independent fraud consultant at newly launched Sumner Fraud Consulting, said the industry needed a more honest conversation about how emerging technologies were changing the fraud landscape and the way investigations were conducted.

“We’re not having that honest conversation about deepfake videos in the new fraud,” she said.

Sumner described a case in which a deepfake video was used to create a fake address linked to an allegedly stolen item.

A field investigator was subsequently sent to the property, only to discover that the location depicted in the footage did not exist.

The investigator instead found a terraced property with no garage and no apparent connection to the reported theft.

“That is the problem that we are trying to tackle, a futuristic problem with really, really old strategies because that’s what we think works,” she said.

Rethinking investigations

Sumner argued that insurers must adapt their investigative approaches to reflect the changing nature of fraud, rather than continuing to apply traditional methods to increasingly sophisticated scams.

“What is the point of knocking on the door of a stranger’s house who has no idea that there’s a deepfake video of their garage having something stolen,” she said.

She added that investigators needed to “educate ourselves to investigate them on their territory better”, rather than relying on existing techniques.

“That, for me, is the biggest challenge,” she said.

The panel also warned that artificial intelligence (AI) could increase both the scale and sophistication of fraud attempts.

Tim Burton, chief product and success officer at Smartnumbers, said criminals were already testing AI-driven attacks against insurers’ agentic chatbots.

“We’re seeing criminals testing that now, getting very good at it,” he said.

Burton argued that insurers needed to focus on how technology was reshaping fraud risks, rather than simply layering AI onto existing processes.

He added that firms required a “consistent understanding” of how customers were identified, authenticated and assessed for risk across different channels as digital and voice interactions increasingly converged.