Artificial Intelligence helps counter cyber fraud
In 2016, North America saw an increase of $16 billion in online financial fraud compared to the previous year. It is no wonder that businesses and organisations are putting great effort into researching the most efficient ways of countering fraud.
Artificial Intelligence is placed quite high on this list.
Artificial intelligence (AI) is the capability of a machine imitating human behavior using data they have collected and reproduce or predict in a highly accurate manner. As machines learn from more and more data, they assess risk profiles and recognise suspicious moves.
AI is a great ally for merchants, especially as credit card fraud rises. Machine learning is able to develop consumer profiles and flag transactions that look suspicious and do not match according information from their database – study on Science Mag has shown that trained computers can pick out liars 90-95% of the time.
AI also aids in decision scoring, as a considerable number of transactions are falsely declined. “We estimate that in the US alone, the value of false declines is more than 13 times the total amount lost to actual card fraud. Applying machine learning to decision scoring is a new way of creating a positive consumer experience, while also minimising fraud.” Says Al Pascual, head of fraud and security at Javelin Strategy & Research
The process of machine learning could be summed up in three phases: the learning phase, the prediction phase and the continuous learning phase.
The learning phase is, once the machine is fed with information, which it explores and scrutinises , then starts deciding which transactions are fraudulent and which are valid. The prediction phase uses the model from the previous phase and applies it to new data. The continuous learning phase is, therefore, when machines are continuously filled with new unprocessed data and behavior patterns.
A security system which counts on the support of AI is able to overcome limitations that human power can’t, such as providing a 24/7 working system monitoring all transactions uninterruptedly. Security personnel could then dedicate this time to focusing on innovation and technological development.
The trend is spreading around finance sector as well. Overseas Chinese Bank (OCBC) has confirmed that it is in the testing process of using machine learning in order to build profiles on customer behaviour, which can be use to detect fraud and money laundering.
So far the protocol had been, when a suspicious transaction arises, the compliance team is notified and initiates a research process through newspapers, social media and search engines in order to understand one’s patterns and those of people around them – a task that takes a least one hour. The algorithm does this in less than minute. The AI is able to map suspicious transactions to judge if they are fraudulent or illegal.
Passwords are outdated, users want biometric authentication
In 2016, the number of biometric mobile payments was just over 600 million. By the end of 2017, this number is set to triple. The increase was initially boosted by Apple Pay, and rapidly followed by adoption of other mobile wallets such as AndroidPay and GooglePay.
A research from Juniper revealed that smartphones with fingerprint recognition also played a big part in opening the doors for widespread use of mobile payments and banking apps. Biometric authentication is indeed a great tool in the payments industry – it is convenient and quick for users and regarded as a safe method of verification.
In times when cases of data theft and identity fraud grow exponentially year after year, biometrics is positioned as one of the best alternatives in the market for ensuring consumer security.
“Given this, it’s unsurprising that organisations are looking to biometrics to help overcome this problem. If an organisation finds it has two different applications, in two different names, but both accompanied by the same fingerprint (for example), it can easily see that at least one of those applications is highly likely to be fraud” said Sandra Peaston from Cifas, UK’s largest fraud database organisation.
A report from EyeVerify has recently provided highly insightful data on how users feel in regards to biometrics and security. Out of more than 1,000 respondents, 82% believe biometric verification is more secure than passwords and 79% said they want more biometrics options than just fingerprint recognition to access banks and payment apps.
Furthermore, 42% of respondents said they would not consider a banking or payment which does not offer biometric authentication. And 82% said they perceive banks whose apps offer biometric verification as being more proactive around security issues.
“Most people use some form of biometrics every day, but they want more opportunities to use it to make their lives easier and more secure,” said Toby Rush, CEO and founder of EyeVerify. “Banks and payment providers have a huge opportunity right now to build brand trust by giving customers the user experience they want. Those who are investing in biometrics authentication will be able to increase customer usage and reduce risk at the same time.”
For merchants, biometrics can increase considerably consumer engagement considering how quickly they can access the platform, as 86% of surveyed agreed that biometrics makes logging into apps easier than entering a password.
Apart from fingerprint, other types of biometric information are being tested, such as face recognition through ‘selfies’, eye and voice.
The industry is confident of the potential of biometrics. BBVA estimates that by 2020, 65% of mobile commerce transactions will be authenticated using biometric technology and the global mobile biometric market will reach $22 billion by the same year.