玩偶姐姐

Our imperfections, more often than not, are the Achilles’ heel for deepfake technology.

We regularly see fake IDs with synthetic faces and data during identity verification. They look close to legitimate, but we see flaws or a lack of them, such as too much symmetry in the portrait or not enough makeup, because it takes a long time to make them look like the real thing.

Deepfakes today don’t get our imperfections right because bad actors don’t put enough computational horsepower into making them. If they did, it would take a hundred times longer. It’s just not economically feasible to make more realistic forgeries.

Quantum computing has the power to change that dynamic. The technology’s speed in certain applications sounds like science fiction. , for example, performed in less than five minutes a computation that would take today’s supercomputers 10 septillion years.

If bad actors enabled their deepfake engines with quantum computing suited for the task, the results could have 1,000 times better fidelity. It’s a scary thought. The face would be slightly asymmetric, the wrinkles would go in the correct direction, a few hairs would stick up and the background would perfectly match that of a department of motor vehicles.

The only way to detect that advanced type of synthetic ID would be to use quantum computing to develop machine learning that can spot the tiny signals in the noise of a fake document. In other words, we’d have to use a quantum computing defensive capability to match the quantum computing attack. 

Preparing for the quantum horizon

Quantum computing isn’t readily available, but it will be in the next several years. Now is the time to prepare for its benefits and risks.

It likely will arrive in the same way as dedicated graphics GPUs in gaming about 10 years ago. Companies at the time wanted to make gaming faster and more realistic by going beyond generic CPU processors.

When they supercharged the chips, they revolutionized graphics and gaming. But the chips had other high-speed applications, such as for medical or weather computations. They also caught the attention of bit miners, who saw the chips as powerful math engines that would be hundreds of times more efficient.

Those graphics chips opened new worlds, including deepfakes, that people didn’t originally envision. The same thing will happen with quantum computing, possibly in the form of a graphics card for gaming.

That card could make gaming even more fluid. A gamer could move through a landscape that looks like reality, with trees, lighting, shade and tall grass swaying in the breeze.

That’s just one quantum computing application. But once the technology is available, bad actors will subvert it.

A new era of high-speed attacks

Businesses today commonly use and to encrypt and store sensitive data, including biometrics, personally identifiable information and passwords. The security is based on the idea that there’s no way criminals could crack it in a reasonable amount of time with today’s readily available computational power.

Quantum computing undercuts that theory with the speed to circumvent a lot of the security built into systems. That makes cryptography and passwords early targets for the technology because it can make so many more guesses per unit of time.

If it took years to crack a password before, quantum computing could do it in seconds. The technology isn’t necessarily smarter, but it’s much faster. 

The promise of quantum computing’s speed in the near future has also created the store-now-decrypt-later phenomenon. Bad actors anticipating a breakthrough in quantum have downloaded encrypted sensitive data that would take hundreds of years to unlock with today’s technology. 

They’re waiting for quantum computing so they can decrypt the data. But they’re downloading it now because they know when quantum comes out, companies will re-encrypt all that data to make it harder for the new technology to access.

It will get to the point that organizations without quantum computing won’t be able to defend against it.

Matching speed with speed

Certainly, if no one changes RSA and ECC, someone using quantum computing could quickly crack them. But businesses won’t stand idle on the sidelines while bad actors arm themselves with the next generation of computing speed. 

They’ll fight fire with fire in security. The same techniques used today to encrypt will be orders of magnitude better with quantum computing. Quantum-enhanced authentication will also strengthen the current methodology.

When deepfakes emerged, AI and machine learning in identity verification found the technology’s weaknesses. The same will be true for quantum computing.

This article is published as part of the Foundry Expert Contributor Network.
Want to join?