Are Deepfakes Really a Threat for Insurers?
Insurance businesses, along with lots of other industries, are becoming increasingly aware of the potential threat posed by the emergence of deepfake media. Deepfakes are highly sophisticated fake images, video or audio created with the help of AI to be almost indistinguishable from the real thing. The emergence of deepfake technology has had implications for a wide range of settings.
This technology has made headlines for the ability to create fake images of celebrities doing and saying things that may be out of character or sensationalist. In more sinister cases these hoaxes can be highly damaging or distressing for the individuals concerned.
Increasingly, deepfakes are being used for fraudulent and criminal activity, so multiple industries, including corporate and commercial insurance organisations, need to be alert to the risks.
Potential for fraud
For the Property & Casualty insurance industry, deepfakes pose a clear opportunity for fraud. At the point of taking out a policy, fraudsters can create documents and photographs of a property or business that does not exist; or amend images to insure a property that may exist, but in an entirely different form. Another, similar type of edited media, known as “shallowfake” uses software (rather than AI) to alter images, often on a large scale. It was reported that one piece of Canadian ID - a passport – was submitted around 2,500 times within a month, each time with changes to the name, address or photograph.[i]
In terms of claims, accidents, fires and vandalism can be digitally hoaxed with very little technological expertise. A quick peruse of app stores reveals a variety of different apps that will change a photo of any car so that it appears to have been “wrecked” as a “prank.” These easy-to-use services can be impossible to identify as fake with the naked eye. As automated services within the insurance industry continue to grow and people are able to make claims with photographs via a computer, it is easy to see where fraud could become prevalent. Where deepfake detectors do not exist, there is the potential for huge losses through insurance fraud.
The current response
A 2022 Study by Attenstiv surveyed insurance industry professionals and discovered that while 80% felt concerned about the potential for fraudulent images to be used in insurance transactions, only 20% of the respondents had taken action against any digitally altered images.[ii]
The insurance sector is increasing automated services at a rapid rate, but the speed that they are innovating is not necessarily keeping up with the rate of fraud detection. In the Attenstiv survey, it is suggested that more than 50% of insurance professionals plan to automate services within a year, while a mere 22% are utilising a form of validation or fraud prevention for digital media. This is alarming, as it highlights a significant gap which fraudsters may be able to leverage. [iii]
Insurtech
The main option for insurance companies wishing to protect themselves, is via Insurtech. New technology is the only real way to combat the rise of deepfake technology. There have been numerous new computer programs developed by Insurtech companies, as well as a number of emerging options.
Attestiv and other insurtech companies, use a technology known as blockchain to verify images and other media. Blockchain is a system by which information is recorded system wide, making it then almost impossible for the image to be altered digitally without being recognised.[iv]
In the UK, an Insurtech firm, BlockFrauds is looking at innovative technologies that can help to combat insurance fraud. They use blockchain technology to provide Claim Handlers with information around the likelihood that a claim may be fraudulent, allowing them to focus on investigating the most likely frauds, and quickly pay out on the non-fraudulent claims. BlockFrauds also works on audio files, with analysis of voice files to assess how likely they are to be fraudulent and producing a “sentiment analysis” for Claims Handlers. [v]
Another type of insurtech, which can be used together with blockchain, is AI claims analysis. One company offering this is Prague based Resistant AI, which analyses each piece of media in over 500 ways, as well as looking at the behaviour of the claimant to assess the likelihood of a claim being fraudulent. [vi]
Utilising Employees
There is nothing that can be done to train an employee to visually recognise deepfake images, as they are not detectable to the naked eye. However, the individual brokers, underwriters, claims specialists and other team members are still vital in the battle against fraudulent claims. Sometimes difficult decisions need to be made around whether to approve further investigation or allow a claim to go through. Sensible conversations need to be had with clients, with human responses. Customer-facing roles are more important than ever, as genuine policyholders must not be inconvenienced by the fraudulent activity of others.
Investing in excellent team members to support the business as it tackles these emerging threats is really important. Specialist insurance recruitment agency, Aston Charles already have many potential candidates who are experienced and skilled at working within Insurtech and understand the risks around deepfake technology.
The Future With Deepfake
Fraudsters will continually look for new ways of defrauding insurance companies, and companies need to react. A combination of Insurtech innovation and skilled, experienced staff can make this a battle that can be won, protecting businesses and supporting genuine policyholders to get their claims processed as swiftly as ever, retaining the reputation of the insurance industry.
[i] Rehak, M. 2022. Forget deepfakes, shallowfakes are the real threat to the insurance industry. 23rd August. Finance Derivative. [Online]. [7 February 2023]. Available from: https://www.financederivative.com/forget-deepfakes-shallowfakes-are-the-real-threat-to-the-insurance-industry/
[ii] Pendleton, B. 2022. New Deepfake Survey Highlights Risks to the Insurance Sector. 19th May. Newswires, EIN. [Online]. [5 February 2023]. Available from: https://www.einnews.com/pr_news/572838813/new-deepfake-survey-highlights-risks-to-the-insurance-sector [iii] Ibid
[iv] Cooper, C. 2022. What insurers need to know about deepfake images. 27th October. Anziif. [Online]. [5 February 2023]. Available from: https://anziif.com/professional-development/the-journal/volume-45/issue-3/what-insurers-need-to-know-about-deepfake-images
[v] Ibid.
[vi] Ibid