Brazilians do not have a minute of peace. This phrase, which became a meme on the internet, seems to have been tailor-made when it comes to fighting fraud. Proof of this is a survey conducted by Serasa Experian, which shows that in Brazil, every 8 seconds a fraud attempt is made. Literally, Brazilians don’t have a minute of peace.
By the way, in this sense, it is possible to refer to another joke that rolls on the internet and says that ‘the scam is there – and it falls even those who don’t want it to! Between June 2020 and June 2021, six out of ten internet users have suffered some kind of financial fraud, according to research done by the National Confederation of Store Managers and Credit Protection Service of Brazil (CNDL/SPC). In amount, this represents a loss of R$ 2.7 billion, including expenses in the search for redressing the problem.
The fight against fraud is an even greater challenge when related to the financial universe. The expansion of fintech and digitalized banking services has proven to be a breeding ground for criminals because of the growing number of users and security flaws in the technology of applications and systems.
Below, understand more about this scenario, the challenges to be overcome in the fight against fraud and how the use of Artificial Intelligence can contribute to solving this problem.
Great opportunities bring great challenges
The technological revolution represents a sea of possibilities for the financial sector. And you don’t need to go far to realize this, all you have to do is analyse yourself: how many times have you consulted your bank statement on your mobile phone? How often do you use your bank’s app to make a PIX? How many online purchases have you made recently? How many times have you used the chat available on a financial institution’s app to solve doubts or problems?
It is no coincidence that the reflexes of this digitalisation of banking operations can be seen in people’s daily lives and draw attention when they are counted in an amount, such as the one raised by a survey conducted by Febraban, which revealed that mobile phone transactions account for 51% of all banking operations. And that PIX, means of instant and free electronic payment offered by the Central Bank of Brazil to individuals and companies, already has 96.8 million registered accounts.
And great opportunities bring great challenges. As people have become more trusting and more frequent users of technology as an intermediary in banking processes, criminals have taken advantage of this.
In this sense, finding ways to mitigate the risks of scams becomes a challenge both for companies, which can suffer great financial and reputational damage, and for consumers, who risk daily having their data stolen and losing money.
Who I am, what I own and what I know
Generally speaking, a good security system should protect consumers in three spheres: who I am – which has biometrics or personal verification as resource examples; what I own – which encompasses mobile phones and digital tokens; and what I know – which is the password.
However, achieving efficiency in the implementation of these three barriers is not one of the simplest tasks. So much so that many banks, fintech and financial institutions adopt the premise that this responsibility lies with the consumer, since the phone is for personal use and must be protected by its owner. There are even legal decisions in favor of the bank, releasing the reimbursement of fraud committed after the device is stolen.
At this point of the text, you may be thinking: ‘if someone steals my mobile phone, there is no problem. The bank’s app only opens with your fingerprints’. What few people know is that circumventing this step is simpler than it seems.
This is because most banking applications use a simplified authentication through the use of biometrics, which means that the fingerprint used for access is compared to the one registered in the device and not in the institution’s system.
In practice, fraud can occur in this way: a criminal – or even someone close and malicious – can get the phone working and already unlocked (as is the case of the gang that stole devices from drivers who were using the GPS locomotion system in São Paulo), register a new biometric on the smartphone and then log into the bank’s app.
Artificial Intelligence in the fight against fraud
Artificial Intelligence has proven to be one of the most effective tools in the fight against fraud. Simulating human intelligence, but with the capacity to analyse large volumes of data in a short period of time, AI is able to identify patterns, cross-reference information, diagnose the behaviour of a user and draw up a profile capable of preventing fraud, blocking malicious actions.
In practical situations, it is possible to perform predictive analysis and tell whether an e-commerce purchase attempt is legitimate or a fraud attempt. For example: before releasing a purchase, in a matter of seconds the system can cross-reference user consumption data, such as frequency, types of products, amounts spent on each purchase, gadgets used and payment methods, to determine the authenticity of a transaction.
In this context, it is machine learning that allows AI to use previously acquired knowledge to promote the security of system users.
Another front of performance of this technology can be related to facial recognition, derived from the combination with computer vision.
In practice, the system that brings together Artificial Intelligence and computer vision can detect information from images that reveal biometric points on users’ faces, interpret this data and make decisions.
Not without reason, the forecast is that in the future – and it is not so far away – Artificial Intelligence will be the major responsible for the control of indispensable processes for the security of an organization, its operations and its customers.
If you are interested in the topic and want to better understand how Artificial Intelligence can prevent fraud, preserve reputation and create solid foundations for the growth of your business, please contact Monitora.
