Quantum computers will soon be here – and this gives rise to concern among several experts. Quantum computers, you see, may potentially solve mathematical problems in seconds that it would take ordinary computers billions of years to handle. One example is breaking the sort of encryption that now protects our private communication, not to mention monetary transactions. Your privacy and your life savings may be at risk, so should you worry?

# Should we worry about quantum computers?

In May 2018, the director of Microsoft’s research unit, Arvind Krishna, went onstage at the Churchill Club in San Francisco and warned the audience that in a little more than five years quantum computers would become able to break the encryption codes we use today. This will mean that our encrypted communication and information will be laid bare and not least that our digital savings and transactions could be hacked in seconds. In theory, the entire foundation of our economy and hence our civilisation could crumble. Does this mean that we should start to panic? Not necessarily, but then again...

As the name suggests, quantum computers have something to do with quantum mechanics: the difficult to grasp natural laws that describe the universe’s most elementary building blocks, atoms and the elementary particles they consist of and interact with. According to quantum mechanics, a particle can be in several different states at once until you measure it, upon which the particle immediately assumes one of the possible states. For example, many particles have a *spin state*, which is left-handed and right-handed at the same time, until you measure the particle. Hence, where a bit in an ordinary computer has the value 0 or 1, a quantum-mechanical *quantum bit* or ‘qubit’ has the values 0 and 1 at one at the same time. Several particles may be ‘entangled’ through quantum mechanics (*quantum entanglement*) and form a chain of cubits that act as a unit. Where a chain of eight ordinary bits can represent one of the values from 0 to 255, a chain of eight qubits can represent *all* the values from 0 to 255 at the same time, since each qubit has two values at once. This makes it possible to make calculations for all these values at once, for example to see if any of the values are prime factors of a larger number. Prime factors are the prime numbers that a larger number is the product of; for example, the number 30 consists of the prime factors 2, 3, and 5 (2 x 3 x 5 = 30). Some numbers only consist of very large prime factors: the number 4,071,461 e.g. has the prime factors 1,949 and 2,089.

## Mathematics

This isn’t meaningless mathematics, as most encryption today uses numbers that are the product of two very large prime numbers, often consisting of several hundred digits. To code a message, you just need the product (the public key), but to decode the message, you need to know the two prime factors that make up the public key.

In theory, you could find these two prime numbers by going through all prime numbers up to the square root of the public key and see if they are a whole fraction of the key, but because the numbers are as great as they are, it would take an ordinary computer a very long time to solve this task. A few years ago, for instance, scientists spent two years finding the prime factors of a number with 232 digits, using hundreds of computers running in parallel.

The above also illustrates the limitations of quantum computers. They are excellent at solving the same problem for many different numbers at once, but if they have to solve different problems one after the other, they are no faster than ordinary computers.

Quantum computers can be used for much else than breaking encryption, so there are good reasons to develop them. Among other things, they can sift through vast amounts of data rapidly to find items of relevance, and this may lead to far faster search engines capable of searching much larger parts of the internet. Perhaps more interestingly, researchers believe quantum computers can revolutionise artificial intelligence, among other things by making machine learning far more efficient. More esoterically, but no less importantly, scientists expect to use quantum computers to simulate quantum-mechanical systems and hence achieve better understanding of chemical and physical processes.

## How far have we come?

Quantum computers aren’t simply a theoretical possibility today. Companies and universities have, in large laboratories, created quantum computers with a small number of entangled qubits, and they function as predicted by quantum mechanics. However, none of them are yet able to solve problems that traditional computers can’t solve faster, and the holy grail is thus *quantum supremacy*, meaning a quantum computer that can solve certain tasks faster than ordinary computers. Quantum computers won’t be a challenge to encryption before this is achieved.

It has turned out to be hard to entangle several qubits, and adding an extra qubit becomes harder the more that are already entangled. At the present, Google holds the record with a quantum computer with 72 qubits, but the company Rigetti hopes to build a quantum computer with 128 qubits within a year. This may sound like a lot, but encryption as described above typically uses keys of at least 224 bits and often up to 4,096 bits. Another issue is that the entanglement of particles in the undetermined state that makes them qubits is rather unstable, and the particles easily become uncoupled by even very small, almost undetectable fluctuations of heat or magnetism – something known as *decoherence.* It is necessary to supercool quantum computers to just above absolute zero to maintain entanglement long enough for calculations – and even then, errors are not uncommon.

The issues with quantum computers will likely be solved over time, but it is impossible to say if we will have working quantum computers with thousands of entangled qubits in 5 years, 10 years, or 50 years. We may already have problems with encryption-breaking quantum computers in 5 years, but it will most likely take longer, and in any case, it doesn’t look like quantum computers will be available for common users any time soon. However, this does not mean that we can breathe easy, for it is quite possible that major actors like national government agencies and some tech giants may get access to the technology within a decade – and it would be naïve to believe that none of these will abuse the ability to bypass encryption.

## Tried and tested

Quantum computers, however, need not spell the end of safe encryption. Scientists are busy developing encryption methods that will be immune to quantum computing, meaning that quantum computers will be no better than traditional computers at breaking encryption. As I wrote above, quantum computers are only suitable for solving certain types of problems, where the same operation is performed for many numbers at once, so it is simply a matter of finding encryption methods that can’t be solved this way.

We have known encryption methods that are quantum resistant since the 1970s, long before anybody talked about quantum computers. However, these methods are far slower and less efficient than the types of encryption we use today. A recent promising candidate for quantum-proof encryption is so-called *lattice-based cryptography*, which is potentially faster in use than current prime-based methods. This does not necessarily mean that this method will replace current encryption methods any day soon. After all, it is easier to use a tried and tested method than experimenting with a new technique with inevitable growing pains, and security actors can’t justify using an algorithm that contains even the least risk of errors or weaknesses. So, most of us can’t expect to get quantum-proof encryption until quantum computers become a real problem. Organisations with particularly sensitive data to protect will likely blaze the trails, among them banks and intelligence agencies, and not until these organisations are satisfied with the security of these new methods will they find common use. Until then, we can just cross our fingers and hope that nobody with quantum computers will find it very interesting to hack our accounts.

## Other interesting articles

## Should we worry about quantum computers?

Quantum computers will soon be here – and this gives rise to concern among several experts. Quantum computers, you see, may potentially solve mathematical problems in seconds that it would take ordinary computers billions of years to handle. One example is breaking the sort of encryption that now protects our private communication, not to mention monetary transactions. Your privacy and your life savings may be at risk, so should you worry?

## Robot utopia or robot dystopia?

We are at the leading edge of a massive wave of automation. Human labour is being replaced by robots and computers on a large scale, and machine labour becomes increasingly competitive. If the trend continues, most of the work we do today – and much else – will be done by machines. The question is what sort of future this will give rise to.

## Neurodating

In the future, you will not have to go through the trouble of finding a partner yourself: that’s what algorithms are for. Researchers are working on a range of methods and technologies that can measure if you and your date are on the same wavelength – literally – by way of biometrical data about your body and brain. The big question will be where we draw the boundaries for what the predictive algorithm of love should decide.

Posted Jun 07, 2019 in Technology Article from Scenario 06:2018