Fraud Decisioning Adapts To Shifting Conditions

Fraud Decisioning Adapts To Shifting Conditions

Just when you think there are no surprises left, this statistic drops: 60 percent of organizations recently polled had no idea their customer data had been breached, because their payments system and process is too opaque to detect it. That’s surprising in 2020 – or maybe “shocking” better captures the vibe.

The world of cybercrime is a veritable dunking in gloomy stats like that one, but the industry isn’t taking it lying down. Advances in fraud decisioning have cybercrooks scrambling into the new arcana of synthetic identities and fiendish mass bot attacks, looking for any possible opening to exploit.

Some industries are more prone than others, as we discovered in the latest Fraud Decisioning Playbook, a PYMNTS and Simility collaboration. With 2020 on track to be a banner year for internet villains, there’s also good news for merchants and brands. Some of the most effectual uses of artificial intelligence (AI) and machine learning (ML) are happening in the anti-fraud space. That also makes 2020 a potential turning point in the battle between legitimate eCommerce and its shadowy foes.

The Issue With Trust

Trust is untrustworthy. It’s a strangely ironic truth in this time of rampant cybertheft. Businesses that function on trust were among the first to be targeted by internet crime syndicates – especially online marketplaces, where there is no face-to-face or voice communication between sellers and buyers.

The digitized trust that enables the payments world to operate doesn’t just backfire on Amazon and Alibaba. In fact, the world’s smartest companies are getting ripped off big-time, all based on trust.

The publishing and advertising industries spent billions creating a maze of ad networks, pixels and click-tracking technology as newspapers died and social media took over. Too bad so many investigative journalists got canned – they might’ve picked up on cybercriminals using digital ad networks like their private ATMs. So much money has been removed from the global online advertising ecosystem that the best we can do is estimate a wild range: In 2019, cybercrooks made off with somewhere between $6 billion and $42 billion. In a word, shocking. The FBI even had a member of the so-called “Methbot” hacker ring extradited to the U.S. for trial – but only after his group stole an estimated $30 million.

This is where the powerful fraud decisioning systems are making a difference, with AI-powered adaptive detection and flexible data ingression that looks for fraud patterns in unexpected new ways.

Safety Is Multilayered

While ingenious new forms of fraud detection – like analyzing mouse movements for signs of humanness, and the biometric detection of deep fakes and synthetic IDs – are pulling ahead of fraudsters, other methods are proving effective. What’s emerging is a multi-point approach to fraud detection that uses AI and ML to analyze torrents of web traffic, while human experts make the tough decisions.

There is support for ideas like adding some “friendly friction” back into transactions, as well as more strident methods like strong customer authentication (SCA). Anti-fraud efforts are increasingly a multi-pronged affair, where overlapping procedures and technologies sniff out the fakes and let the “real you” get through.