This insights article is by David Janson, VP EMEA, Fivecast
Insurance fraud continues to escalate from opportunistic claims to organised crime networks, costing the industry billions each year. Traditional investigative approaches simply cannot keep pace with the scale, speed and sophistication of today’s digital footprint. Given the magnitude of the challenge, AI‑enabled open‑source intelligence (OSINT) has become essential for fraud teams who need to see more, surface risk faster and make confident decisions.
The Association of British Insurers (ABI) reports that detected insurance fraud rose to £1.16bn in 2024, up from £1.14bn in 2023, and that figure represents only what investigators uncovered. Beneath the surface, billions more in undetected fraud continue to distort risk, erode trust and push premiums higher.
The insurance industry needs confidence that the hundreds of millions invested in fraud detection are being channelled into the most effective, high-impact solutions. Social media, spanning both major and niche platforms, presents a significant, yet underutilised, opportunity. With enhanced AI-powered capabilities, insurers can streamline detection, generate actionable intelligence and achieve measurable performance advances alongside strong returns on investment.
Social media has become a critical environment for identifying both opportunistic and organised fraud. However, the sheer scale and complexity of this data landscape mean investigators are currently only scratching the surface of its potential as a rich open-source intelligence resource.
Mapping connections
Advances in AI-driven open-source intelligence (OSINT) are transforming investigators’ ability to handle the vast scale and complexity of digital data. Modern analytical capabilities now allow OSINT teams to rapidly surface relevant signals, enrich insights from diverse sources and highlight indicators of risk up to 400 times faster than manual methods. These powerful and cost-effective tools enable investigators to identify emerging patterns and suspicious behaviours linked to potential fraud.
Social media insights reveal hidden fraudulent behaviour.
Investigators need advanced tools in order to fully interpret and analyse social media, which has become a powerful source of intelligence for fraud detection. This year, for example, a woman failed in her £400k injury claim after being spotted playing rugby, walking muscular dogs and sharing posts on social media that contradicted her case. We are all familiar with stories of fraudulent injury claimants who post about themselves working out in the gym or participating in other physical challenges.
From crash-for-cash scams, ghost-broking and shared tips on how to defraud flood damage or holiday theft claims, the perpetrators and their affiliates leave a digital footprint that AI-powered OSINT solutions can equip investigators to identify and analyse in minutes.
Overcoming data challenges
The ability to analyse millions of social media posts at speed and correlate findings with other databases produces a level of insight that humans trawling through the internet can never attain.
Statista estimates there are nearly 55 million social media users in the UK, with the number expected to reach 65 million by 2027. Given this scale, it is impossible for investigators to manually review every mainstream platform – let alone navigate the deep and dark web where criminals recruit, coordinate, and exchange tactics. The sheer volume and complexity of this data demand advanced, AI-powered tools that can surface relevant signals at speed and provide investigators with clear, actionable intelligence.
Augmenting human expertise
The argument is not that investigatory expertise should be replaced by technology, but rather that it should be vastly augmented by it. Teams investigating insurance fraud need to combine their investigative skills with OSINT technology that automates data collection, quickly filters and correlates massive amounts of multimedia data to detect risks.
In an insurance context, OSINT solutions integrate external intelligence sources, including registers of known insurance offenders. The final decisions, however, are always made by investigators or assessors. This is the correct approach to tackling today’s insurance fraud without relying entirely on increased headcount or the refinement of outdated manual processes.
Avoiding excessive intrusion and surveillance
OSINT technology can provide vital insights by applying configurable data analysis. Modern solutions are user-friendly and do not require specialist data expertise. Even so, they go beyond keyword searches by analysing images, videos, memes, posts and other text-based content. Where the main source behind a piece of content is not immediately clear, analytics can often identify linked accounts and related online connections.
All this can be achieved in full compliance with robust ethical frameworks, privacy and data protection regulations. While efficiency is the objective, automated OSINT that follows these guidelines avoids collecting irrelevant or excessively intrusive personal data and is compliant with the GDPR and the EU AI Act. Insurance fraud investigation needs responsible AI with a high level of transparency, applied under shared standards that ensure proportionate practices.
This is the only way to accelerate investigations, reduce the scale of fraud and reduce costs. For whatever reason, many members of the public are prepared to collude with criminals to defraud insurers, while others are sucked into online scams such as ghost-broking. To tackle all these types of fraud, the insurance industry should be leveraging greater automation and AI-powered OSINT technology.
Be the first to comment