By John Kipp, Sera-Brynn COO
As advancements in Quantum Computing and Artificial Intelligence are always popular media topics, I thought it’d be interesting to take a look at how the two are converging and what such a convergence might mean for cybersecurity.
First things first though. Below is a quick and very basic primer on Quantum Computing and Artificial Intelligence:
Today’s traditional computers operate by processing information as “bits.” A bit can be either a “1” or a “0.” Even massive data-crunching super computers such as the US Department of Energy’s Titan Cray XK7 operate in this fashion. These computers explore potential solutions to problems in a methodical, sequential manner.
Quantum Computers on the other hand, store data as a “1,” a “0,” or both at the same time. Accordingly, they do not use bits. In Quantum Computing, they’re referred to as a “qubits” (quantum bits). There is more data per qubit than there is per bit, thus allowing for exponentially faster calculations. To add to it, Quantum Computers can explore potential solutions to problems simultaneously versus sequentially.
The ability to perform simultaneous computations is known as quantum parallelism. Here’s an example of how this differs from conventional computers: If you have ten million random numbers and need to find a specific one, a conventional computer will have to search methodically, one number after another until it finds it. A Quantum Computer, depending on its specs, could search thousands of numbers at the same time.
This all sounds very promising, but one thing to keep in mind that this is a very new technology. At the moment, Quantum Computers are not ready to replace conventional computers. For starters, Quantum Computers excel at tackling a few specific types of problems such as factorization, certain types of simulations/experiments, and “needle in a haystack” problems such as searching through massive amounts of data. Also, programming languages and software interfaces are nowhere near the level of maturity they are for conventional computing architecture.
Artificial Intelligence (AI)
“AI is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable.”
– John McCarthy, Stanford Computer Science Department, 2007-11-12
A lot has changed since John McCarthy coined the phrase “Artificial Intelligence” in 1956. Today’s examples of AI really do seem like something out of science fiction movies with self-driving cars and the world’s best poker players being beaten by a computer. In recent years, AI has also defeated the world’s best in Chess, Go, and Jeopardy.
AI is showing up right in our homes now too with Amazon’s Alexa, Google’s Google Home and Microsoft’s Cortana AI butlers. New ones such as Samsung’s Bixby (formerly Viv) are showing even greater capabilities by allowing users to continue a line of questions within context. Example: You might ask “When is the Super bowl going to be on TV?” It will give you the answer and then you might ask “What will the weather be like that day?” and it will understand that this question is a follow-up (proper context) and give you the right answer.
Next up will be driverless trucks, pilotless passenger and cargo airplanes and even space exploration. There is already some rudimentary AI in use in space exploration. For example, the Mars rover Curiosity has a system called AEGIS that determines which rocks to analyze with its laser spectrometer. There are many problems dealing with space exploration that AI can help with. One is basic communications between Earth and Mars. It can take nearly half an hour for a message to get from one to the other. If the devices on Mars are capable of making more decisions on their own, countless hours (days? weeks?) will be saved.
Quantum Computing and AI
So what happens when you have Quantum Computing backing an AI? First, it depends on availability of the technology. Right now, the most plausible option for purchasing a Quantum Computer is from a company in Canada called “D-Wave.” It will cost approximately $15M if you want to purchase their 2000Q Quantum Computer. Suffice it to say that this technology will not be on your desktop any time soon, but it is maturing at an incredible pace.
Google’s Quantum A.I. lab is working with D-Wave and NASA to develop solutions to problems that are difficult for traditional computer architecture to deal with. In their own words: “We are particularly interested in applying quantum computing to artificial intelligence and machine learning. This is because many tasks in these areas rely on solving hard optimization problems or performing efficient sampling.”
From a cybersecurity perspective, this is significant. If you have an entity with the resources to obtain or develop a powerful enough Quantum Computer, it may be able to quickly and easily break today’s most prevalent encryption methods. This is something even the NSA is taking very seriously.
Others with a focus on cybersecurity are taking it seriously too. On January 24, 2017, D-Wave informed the world that the first customer to buy their 2000Q Quantum Computer was a U.S.-based cybersecurity company. But beyond simply breaking encryption (which is no small feat), if a machine that powerful, using AI, is able to learn and assess incredibly complex problems in milliseconds, it’s not much of a leap to assume that one day, it may be able to probe, listen to, and attack an organization of any size (or several at once), instantly find weaknesses and patterns, capitalize on them, react to impediments, steal data or cripple infrastructure…all at a pace inconceivable to even a large group of humans.
This is not a problem we’re going to be facing tomorrow since both of these technologies are so new. But years down the road, we may find ourselves dealing with the possibility that we’re in a race to see whose Quantum AI is smarter and faster. Unless we have a self-aware Skynet or HAL9000 situation – and then all bets are off.