How AI can affect NHS Fraud

Hopefully you will now have a better understanding of what AI is. Next, we will look at different areas that it may affect fraud in the NHS. It is important to remember that using AI is not a crime. It can be used by criminals to help commit fraud more effectively.

Application forms

The Counter Fraud Team has been sent numerous dodgy-looking application forms to review. It is suspected that many of the responses on these have been created with the help of AI. Typing “how can I describe how I am a good team player” or similar into a Generative AI app powered app will come up with a beautifully written response. At the time of writing, there is no ‘ban’ on using AI to assist in drafting application forms – it is no different to finding samples from Google or asking a friend for advice. In all of these situations, the crux of the matter is to make sure that what is written is truthful.

AI (like Google and friends) can make things up which do not apply to the applicant, and therefore the application form may be fraudulent if it makes false representations. For example, we have seen somebody giving examples of tasks they had done in their role as a Project Manager – but in their work history, they had never had a job as a Project Manager. By contrast, if a Care Assistant asks AI to “list tasks a Care Assistant does” and the applicant chooses only the responsibilities which actually apply to their role to add onto their application form, it would not be fraud.

Phishing emails

Phishing emails are sent to try and trick the receiver into doing something, such as entering their log in details into a fake website or make a payment which is not required. These emails are often sent from organised crime groups who can be anywhere in the world and poor spelling, terminology and grammar could be indicative that they were not genuine. AI will allow the fraudsters to create personalised and professional sounding emails more effectively.

Invoice fraud

For years, fraudsters have used many techniques to try convincing organisations to make payments for goods and services which have not been received. AI is going to assist the fraudster in many ways, from analysing publicly available information to spot patterns of spend, to drafting realistic looking fake invoices.

CEO Fraud Fraudsters impersonating senior staff may become more difficult to spot if AI is used. At present, this will be as described above, with more convincing emails. As AI technology develops, we anticipate that voice recognition and deepfake videos could be used too. These are just a few ways which AI could be used to assist fraudsters target the NHS. It is important to be aware of this developing risk, but current advice remains the same. If you are concerned that you may have spotted fraudulent activity or you’re unsure about something you have received and need some advice, please contact your Local Counter Fraud Specialist.

This article is in the 'AI Special' June edition of our Counter Fraud Newsletter which can be found her Publications page.