Shallowfakes: the surprisingly simple claims fraud set to rival AI

Shallowfakes are costing insurers serious money and fuelled 2023’s headline-grabbing premium hikes. Not to be mistaken as a lesser threat than deepfakes, shallowfakes are disturbingly effective and growing in prevalence. So, how did Insurance arrive at this point, and what can be done to course-correct? Our digital fraud experts explain below.  

• What are shallowfakes? [Skip ahead]
• Why is shallowfake claims fraud so dangerous? [Skip ahead]
• How did shallowfakes become such a problem in claims fraud? [Skip ahead]
• What is the cost of shallowfake fraud? [Skip ahead]
• What can insurers do differently to combat shallowfake fraud? [Skip ahead]
• What is the future of shallowfakes in claims fraud? [Skip ahead]

What are shallowfakes?

Shallowfakes are fraudulent images, videos, or documents created without Artificial Intelligence (AI). Crucially, any bad actor – casual fraudsters to organised gangs - can create a shallowfake. 

No knowledge of or access to advanced AI technologies is required. As such, there is every possibility that insurers will become inundated with shallowfake fraud.  

Why is shallowfake claims fraud so dangerous?

The shallowfake’s accessibility elevates this fraud typology to urgent threat status. However, the real art of the shallowfake is its subtlety. 

Small manipulations or “vanilla” synthetics are easily missed by the human eye and painstaking to uncover during claims fraud investigations. 

As a result, false claims evidence – whether documents, images, or video – stays under the detection thresholds managed by human investigators or standard technologies. 

This means that many insurers only realise they have been defrauded much later, when considerable investigative resources and expensive legal challenges are inevitable. 

How did shallowfakes become such a problem in claims fraud?

Like many of today’s business pressures, shallowfake proliferation is a symptom of pandemic-era operational shifts.

Pre-2020, insurance digitalisation was in motion. However, necessary remote working accelerated change and brought forth processes which include:

• Touchless transactions 
• Self-service policies 
• Straight through processing 
• Automated claims handling 

These smooth journeys ensured that customers could access services. But the minimal friction had an unintended effect: easy exploitation by insurance criminals - equipped with homemade fake documents. 

What is the cost of shallowfake claims fraud to insurers and consumers?

Today, insurers and their customers are facing somewhat of a reckoning.  

Evasive shallowfakes are bypassing many current fraud prevention measures. The resulting unnecessary payouts are unfairly driving up premiums and, of course, cutting into the bottom line. Even when identified, shallowfakes are spiking referrals and placing undue pressure on those investigating fraudulent claims associated with doctored documents. 

The bottom line? Honest parties pay a high price, while fraudsters profit with relatively little effort. 

What can insurers do differently to reduce shallowfake claims fraud?

To combat shallowfake fraud, something must change in claims fraud prevention. Today, the best solution is one that can:

1. Rapidly handle and make sense of huge amounts of unstructured data. Some estimates suggest that up to 80% of an organisation’s data is unstructured. Therefore, given the formats that fraudulent images and documentation must take, this figure is likely mirrored for data associated with fraud.
2. Analyse the minutiae of digital images, including tiny deviations in a file’s digital footprint. The accessibility of manipulation tools is making shallowfakes more convincing.

At present, Synectics Solutions is piloting an exciting digital object checker tool – ideal for any organisation needing to clamp down on opportunistic and organised shallowfake crime. It combines AI, digital forensics, and syndicated data to detect fraud indicators and provide a risk score for use cases such as:

Clamping down on the above frauds means that insurers can better control the impact of crime on customer premiums and investigative resources, and expedite the handling of valid claims.

What is the future of shallowfakes in claims fraud?

Given that shallowfakes have been so successful, there is little motivation for bad actors to move away from simple production tools – even if generative AI is now widely accessible. 

The key takeaway is that despite being less sophisticated than their deepfake cousins, shallowfakes must not be underestimated. Their ease of access means there is huge potential for shallowfakes to wreak serious havoc, especially via organised crime.  

Synectics Solutions is here to support you in your fight against fraud. Why not talk through your challenges with one of our consultants? Click here to set up a call at your convenience.