Detecting deepfakes or training vigilance has a limited impact on fighting deepfake fraud. But identity platforms can help future-proof us against this threat, writes Daniel Robus, chief revenue officer at Contactable.
In early January, an employee at an undisclosed Asian bank sat on a video conference call with senior managers of their company, including the chief financial officer. During the call, they were instructed to move $25-million to various accounts. It was a scam – the employee was never in contact with their bosses but got fooled by a deepfaked video.
2023 started with a bang as generative artificial intelligence (GenAI) went from obscure innovation to a technology reshaping the world. GenAI can create prose and images that could pass for human-created content.
It can also create deepfakes: convincing versions of real people. Most recently, pop star Taylor Swift was the latest victim of deepfaked pornography, and a robocall featuring a fake voice imitating US President Joe Biden tried to dissuade people from voting.
Neither case is surprising. Back in 2022, two-thirds of surveyed organisations experienced deepfake attacks. But the issue has since scaled up to industrialised proportions and shows no signs of slowing down.
Deepfakes are a real problem, especially for identity. The digital world relies on identity: it determines who or what may access a given service, network connection, or data. Indeed, the real world operates on identity: whether you present your driver’s licence at a fancy club or enter your bank card’s pin, you verify your identity.
Know-your-customer (KYC) and anti-money laundering (AML) systems rely on identity. Companies use identity platforms to ensure they deal with the proper party to avoid fraud and fines. Yet, what happens when you can no longer trust someone at face value? What if the person you are seeing or hearing is not real?
There are proposed solutions to the problem.
Foremost is deepfake detection technology. But, while this solution works in some circumstances, it’s in an arms race against ever-improving deepfake generators.
Regulations can help curb deepfakes, but they have no jurisdiction over rogue nation-states and criminal groups.
The third strategy is to train people to spot deepfakes. But this is not very effective – in studies to test their detection skills, a minority could spot good deepfakes (and the researchers note that these participants expected to see deepfakes).
Yet there is another solution: automated, multi-layer biometric checks and identity intelligence harvested from independent sources. Every layer a deepfake adds makes it more vulnerable to discovery.
Adding fingerprint authentication makes it much harder for criminals to fool others. And include a database that queries hundreds of other identity databases to create granular, data-derived identity flags, done at speed by a specialised AI.
The way to deal with deepfakes is like any facade: try and peek behind the curtains. Look for angles the criminals didn’t cover. Verify before trusting. Of course, we can’t expect every person to have such skills or for every company to create such an elaborate system. But identity platforms can.
Platforms are a type of operating system that power and combine different types of software. You are using a platform if you use Microsoft 365, Google Workplace, Slack, SalesForce, or any new generation online software. For example, Gmail, Google Docs, and Google Meet are part of the larger Google Workspace platform. They integrate with each other, creating a combination of services.
Identity platforms do the same. But, instead of office tools, they focus on identity tools.
Doing so creates an enormous advantage. Platforms have two other traits. They continually update and improve – platform users often get access to the latest updates and new services without any upgrades or new costs.
And platforms use APIs, special software that connect unrelated systems. Platforms deliver the latest technology breakthroughs and enable users to connect the platform to their business systems.
Here’s an example: a car dealership will have sales, logistics, customer management software, and a biometric system for physical access. All four benefit from identity.
But rather than each running an isolated identity system, they connect to an identity platform. They get all the same benefits, the cost to the company is much lower, and any new identity upgrade will go where needed. The identity platform is invested in continually innovating, improving and expanding its services. It’s like having an identity specialist for every business process.
The same identity platform can create a united verification front against deepfakes. Deepfakes are often quick and opportunistic.
Yet a bit of proper scrutiny can expose them, and identity platforms are the ones to do the job. They counter the deepfake arms race by continually improving detection without their customers lifting a finger. They work in the background and can automate identity verification so employees don’t have to. And while laws are slow and broad, identity platforms are quick and surgical.
The platform’s researchers can spot the latest deepfake tactics and implement countermeasures that become available immediately to the platform’s users.
There is no final remedy for deepfakes. We cannot legislate them away, and there is no magic technology wand that will make them disappear. As we enter a future where most online content will be synthetic, the best tactic is to use identity as a countermeasure.
Identity platforms help us weaponise identity technologies against deepfakes and other identity crimes. Even if deepfakes aren’t on your radar, fraud and cybercrime are. They all relate to identity, so make identity the solution with the power of platforms.