Preventing fraud has never been an easy task…is artificial intelligence a modern approach to overcoming this?
We’ve all heard about artificial intelligence (AI) and the wonders that the likes of Google and Tesla are creating with it.
Until just a few years ago self-driving cars were something learner drivers and science fiction fans could only dream of, yet Tesla has successfully put their first models on the market, even if they are a bit on the expensive side.
Making appointments and bookings getting in the way of your busy life? Google Duplex will make them for you over the phone just like a normal human.
And of course, we can’t forget Siri, Apple’s loveable (well most of the time) helping hand who can set reminders and send texts when you aren’t able to yourself.
But AI isn’t just about making our lives easier
The media loves to focus on the amazing ways in which these multibillion dollar companies are revolutionising the way we live our lives. What goes on in the background, however, is arguably more important.
Fraud prevention and detection has always been a struggle and with continuous advancements in the way that transactions take place, it hasn’t been getting any easier.
Security giant McAfee launched an investigation in 2016 which found that cybercrime, much of which is fraud, was costing the world around 0.8% of global GDP.
Thanks to artificial intelligence, this is changing.
Financial institutions and many businesses now use software that tracks all transactions, from credit card payments to invoices, and detect when fraudulent behaviour takes place.
They do this by using two branches of AI, machine learning and deep learning, which involve robots learning by example, allowing them to recognise abnormalities in spending patterns or a transaction process, and then reporting this before anything further can happen to the money.
In the past, forged invoices or stolen card details would often be overlooked until the damage was too much to undo. In the best-case scenario, someone would notice when they received their statement, but even then, they may only be able to cancel their card, preventing future usage but not reimbursing them for their money lost.
Now, however, we have AI, a modern approach to stopping fraud that detects the criminal activity almost instantaneously, meaning the money remains safely in the right hands.
Wouldn’t it be better if fraud was prevented all together?
Yes, of course it would. As technology to help stop fraudsters advances, so too does the technology to aid them, so in reality fraud detection may not be enough.
Financial Fraud Action UK’s Fraud the Facts 2017 found that by using AI to detect fraud, only 64% was prevented. Although this may seem like a substantial amount of money saved, that still leaves £3.60 of every £10 unaccounted for.
So, AI that can prevent fraud before it takes place would be a huge benefit to society, but is it possible for anyone, even artificial intelligence, to stop fraud before knowing if it is taking place?
Maybe fraud detecting AI, in spite of its faults, is the success story we have been looking for.
What does this mean for the future?
The company Rainbird has predicted that, if all UK businesses used AI to detect fraud, collectively they could save £7bn in five years. They have also said that as new AI technology is constantly being developed, the 64% of money saved could easily become 80%.
The hope is that as more and more fraud is detected, the criminals behind it will become less incentivised to attempt to carry out the crime, as the risk of failure is too high.
AI is advancing every day, so more fraud is being detected than ever before, and this will continue to improve, meaning more fraud will be stopped and hopefully less will take place in the first place.
So yes, artificial intelligence is a modern approach to the old fraud problem.
Thea worked at ITESOFT as an intern for a week to learn more about marketing. She has just finished her first year of A Levels and hopes to go to university to study economics.