In October 2024, CSA hosted the 3-day Global AI Symposium. Experts at the forefront of artificial intelligence (AI) delivered insights into the benefits, uses, and misuses of AI.

A standout session was “Quantum Artificial Intelligence: Exploring the Future of Intelligent Computing” with sisters Mehak and Megha Kalsi.

Mehak is an industry leader and consultant in the cybersecurity, quantum computing, and CMMC spaces. She is also a Co-Chair of the CSA Quantum-Safe Security Working Group and a CSA Research Fellow. Mehak holds industry recognized certifications including a CISSP, CISA, and CMMC-CCP along with several professional certifications from Stanford University and MIT in Advanced Cybersecurity, Quantum Computing, and AI respectively. Mehak is an alumnus of the University of Florida with a Bachelor of Science in the heavy sciences and Master of Science in Information Systems. She’s been featured twice by Authority Magazine as a leader in the quantum computing industry.

Megha is a distinguished cybersecurity Partner at AlixPartners with expertise in cybersecurity strategy, emerging technologies, security organization development, and AI Governance and Oversight. She is a trusted advisor to Fortune 100 clients seeking to fortify their digital defenses. Megha holds a number of certifications including a CCISO and CISSP. She has held multiple interim cybersecurity leadership positions in which she has turned around security operations to propel efficiencies. Further, Megha has worked hand in hand with hundreds of companies across a wide variety of industries. Both Mehak and Megha share their expertise with global audiences by authoring thought leadership and speaking at conferences.

Below, read an edited transcript of their informative session from the event. Discover how the cutting-edge fields of quantum computing and AI are shaping the future of intelligent computing. At the end of the article, you’ll find follow-up questions from the session, answered by Mehak and Megha themselves.

We also encourage you to watch the full recording of their session on YouTube.

---

Introduction

Mehak: Thank you for joining today. You may be familiar with the topic of quantum computing and you may be familiar with artificial intelligence. But today we'll be covering the future of quantum artificial intelligence, or QAI for short.

Megha: Hi everybody, my name is Megha. I’m so excited to be here today and speak to you about quantum AI.

Quantum Computing 101

Mehak: At a high level, quantum computing isn't just smaller and faster versions of whatever we have today. It's actually a shift in how we process information. We're moving towards more powerful systems that can handle very complicated computations. This shift is essentially from classical computing to quantum computing.

A classical computer is just anything that isn’t a quantum computer. So your laptops, your desktops, any computing you do on your phone, supercomputers - that's what we consider classical. Everything else is quantum.

To start off, what is the most basic difference between the two? First, classical computers run on bits - zeros and ones - that's about it, you have one of two choices.

Qubits are what quantum computers run on, “quantum bits” for the longer version. Qubits can be one, zero, or just about anything in between. If you think of a qubit as a round sphere, it wouldn't just have an X and a Y axis. You also have to consider a Z that allows it to be just about anything, anywhere on that sphere.

For a qubit to be anything between zero and one is for it to be in superposition. We usually run a Hadamard gate, or an “H gate,” that basically moves these quantum bits into superposition. This is really important because one of the basic levels of quantum computing is the ability for a qubit to be in superposition.

Next, when you run a computation in classical computing, it's going to give you the same response every time. With a quantum computer, it can be different depending on how the qubits function and the type of information you put in. You may get a different result.

That leads us into how processing is done. With a classical computer, you have serial processing. You could run, in all fairness, several computers together and create the parallel behavior that we see with quantum computing. But it's not parallel in the way that a quantum computer is.

This is a really important term too: entanglement. We do this by running something called a CNOT gate. Entanglement is basically creating a relationship between qubits so they can do multiple computations at the same time.

Circuit behavior for classical computing is basically classical physics. For quantum computing, it’s quantum mechanics. Your laptop that you're probably using at this time is running at a reasonable temperature within your environment. It's functioning. A quantum computer runs a lot colder, something close to zero kelvin.

The last bit we're going to talk about is Shor's algorithm. It was created in the 90s by Peter Shor. Right now, you can run it on a classical computer. But it won't do what it needs to do in a reasonable amount of time, which is to break current asymmetric encryption. Asymmetric encryption is the basis of public key encryption, which encrypts about 90% of our data.

The theory is that if you run Shor's algorithm on a quantum computer, in a very reasonable amount of time, that asymmetric encryption will be broken. It's something that we all should be thinking about moving forward.

Artificial Intelligence 101

Megha: So you have probably seen multiple definitions of artificial intelligence. We're going to go through some of them here just to baseline our understanding of artificial intelligence in general.

• First definition, “leverage computers and machines to mimic the problem-solving and decision-making capabilities of a human mind.”
• The second one is just having the ability to perform cognitive functions associated with the human mind.
• Third is that it's a given set of human-defined objectives, able to make predictions, recommendations, and influence real or virtual environments.

One of the things in common with all three of these definitions is some sort of reference to humankind, human behavior, and so on. That is at the core of artificial intelligence. Where we want to get to is to achieve human behavior with artificial intelligence.

There are three different types of artificial intelligence that we want to be aware of for this session. There is artificial narrow intelligence, artificial general intelligence, and artificial super intelligence.

What we've been used to in the past few years is artificial narrow intelligence. This is capable of performing a single limited task at a time. It cannot go across multiple domains. It stays on a specific domain or focuses on a specific context.

It’s the majority of current AI systems that you see today. For example, many of the AI technologies you may have seen, such as image recognition, speech recognition, natural language models, or natural language processing, fall under artificial narrow intelligence.

What companies are trying to get to right now is artificial general intelligence, which is AGI. It strives to achieve human-like cognitive ability. We want AI to become more human-like. It is capable of learning and applying knowledge across various domains.

What you're seeing with OpenAI, DeepMind, and so on, is that we’re moving towards artificial general intelligence and attracting very significant attention.

Where we ultimately want to get in society is to artificial super intelligence, which is surpassing human intelligence. We want artificial intelligence to have emotions, beliefs, decision-making, processing, very high IQ systems, or whatever these are really going to be. This is cutting edge research.

To get to super intelligence, we need some additional technology such as quantum computing. But we also need to have a better understanding of artificial intelligence, ethics, accountability, and trustworthiness. To get there, there definitely needs to be more guardrails and more legislation.

How AI Can Help Quantum

Mehak: How can AI help quantum computers be cryptographically relevant? One of the things that quantum computers have a problem with today is noise reduction. Noise, in the context of quantum computers, is basically any type of noise, vibration, sound, or even fluctuations in power, that can affect a qubit and cause an error.

AI could potentially come in and help figure out a way to reduce noise and prevent qubits from decohering. Decoherence is basically a qubit falling apart - falling out of its quantum state and not being able to function as a qubit and do its job. It's important and there are organizations that are working on this right now.

The other part is algorithms. We've talked about Shor’s algorithm, which would take the public key encryption that most organizations use and make it irrelevant. We want to use AI to create quantum-safe algorithms. Now are there some concerns with that? Yes. We will want to do this very carefully.

We are still trying to see a quantum advantage. A quantum advantage is when a quantum computer will be able to do a calculation faster than a classical computer. We are yet to get there, but we are on our way. We're hoping to simulate quantum systems on classical systems, but use AI and figure out what we can do better to get to this quantum advantage.

How Quantum Can Help AI

Megha: The beauty of both of these technologies is they can help each other. Next, I’ll talk about quantum computer uses for AI and how to expedite AI technology.

Currently, a lot of people are struggling with cleansing data. I hear things from my clients such as, “I have so much data, I don’t know how to clean the data. I don’t know how to get data that is good quality.”

Quantum computing going mainstream would help us more efficiently and effectively clean data before we teach AI. That's going to result in the improved learning process of AI. As AI models are training and learning and the test data is becoming cleaner, the learning process is going to get better. AI is going to become more accurate and effective.

The second one is NP-hard problems. NP-hard problems are basically very complex mathematical problems. They’re very time-consuming to calculate using classical computers.

But with the implementation and inclusion of QAI, you're going to be able to do these very complex computations quickly, sometimes within minutes or seconds. Imagine what that can impact. That can impact logistics, scheduling, drug discovery, any sort of code analysis.

And then quantum neural networks. Right now, AI has normal neural networks on classical computers. With quantum computers, you're going to have neural networks that implement quantum principles. Things are going to be faster when it comes to fast learning. It’s going to result in a more robust model that will accelerate matrix operations and deep learning algorithms.

Quantum AI Use Cases

Megha: With that being said, let's move on to case studies and use cases of QAI. The first is one of my favorites, which is augmenting cybersecurity. When quantum and AI come together, it is going to cause this wonderful explosion in the cybersecurity industry, with cybersecurity tools and technologies that will allow us to identify threats faster.

For example, quantum-enhanced AI for threat detection is going to occur. That will allow us to use quantum machine learning to uncover patterns and correlations in cyber-attack behavior that we can't do in real time right now. So, our threat detection will get better, more sophisticated, and zero days will be a lot easier to tackle.

Another example is AI-driven quantum cryptography management. So for example, you will be able to use different types of data encryption to ensure that all your sensitive data is encrypted in the way that it needs to be. And it's going to be done better and easier.

From a quantum AI for advanced risk analysis perspective, right now our risk analysis is slower than what we're going to be able to do in the future. Quantum computing is going to supercharge AI's ability to perform risk analysis and identify complex cyber-attacks. It'll allow us to actually identify some of the more obscure vulnerabilities and prevent them from being exploited. So you'll see a lot of really good things coming together, even for the cybersecurity industry.

Moving on, the next one is global supremacy. The intersection of quantum and AI is going to impact the global economy with impacts on industries such as finance, healthcare, energy, and technology. It's also going to impact the military, warfare, cryptography, and code breaking capabilities.

They will also have very positive impacts, in regards to helping prevent climate change and minimizing environmental impacts and so on.

Quantum and AI is going to identify the most influential global player in the world. That's why you see a competition of countries around the world trying to get to quantum and AI faster than the other. It's really going to impact economies.

Mehak: The last use case we'll touch on today is autonomous vehicles. Imagine autonomous vehicles processing sensor data through quantum computers. They have an AI that will take that information and process it on a quantum computer. It will be able to perceive the environment around the autonomous vehicle better and faster. You'll have AI doing a little more of the thinking than what the sensors can actually do today.

And actually running the AI on the quantum computer will allow it to think faster. You'll get that real-time processing of the data and faster decision-making for the AI. You'd actually be able to have truly autonomous vehicles.

But what really interests me is that you can use QAI to improve the autonomous vehicles themselves. Right now, the batteries that are used in electric vehicles and other autonomous vehicles tend to usually be electric. The batteries are expensive. They use very precious materials that eventually we're going to run out of.

In order to move forward with autonomous vehicles, and electric vehicles in general, it'll be really helpful to use QAI to identify different materials to create these batteries and potentially make the batteries run on less power.

So again, an environmental impact comes in here, not just from those materials that we want to get to be better, but also maybe using less of them, having the vehicles run for longer on less power, things like that.

Quantum AI Ethics

Mehak: Now we'll touch on ethics at a high level. There are ethical concerns for the different technologies independently. And then for quantum AI, you're going to see a lot of overlap in addition to their differences.

Of course, with artificial intelligence and QAI, bias is huge. It's always going to be number one. There are issues with AI being trained on data that isn't considering all different genders, races, economic situations, age groups, or even ableism to a certain degree. There's a lot of bias that we need to move past. We need to make sure that the data we're training our AI on is accurate and reflective of everybody.

If we're going to be using AI or quantum AI in criminal justice or healthcare, it's really important that the data we're training the AI on is accurate, but also that there's some guardrails around it. We need to make sure that we train our models properly and that there are still humans who are helping make these decisions.

There's also the weaponization of artificial intelligence, and even quantum computing to a certain degree. We need to really keep an eye on using it in warfare.

Consider also the autonomy of artificial intelligence and transparency. A lot of times with AI, when it does a calculation or does some amount of work, it's in a black box and we don't know how it came up with its answer. How did it say, “Hey, this is the right thing we need to do here.” That's part of the ethical concern. We need to figure out how it thinks.

Now, moving towards quantum computing, number one is data security. How do we make sure we're ready for our future when that's such a huge implication of quantum computing?

Next is the environmental impact. It's expensive to cool a quantum computer right now, and it's expensive to run artificial intelligence and quantum computing. How do we start minimizing that? Maybe we can use these technologies to help us do it.

Unforeseen consequences are an interesting case. This is with any technology, right? Any emerging tech. The internet was considered emerging tech at some point and so was social media. Both have had unforeseen consequences that we weren’t able to predict and this pattern of unpredictability will continue. We should prepare as much as we can, but we need to keep that in mind and keep our eyes open for unforeseen consequences.

Quantum AI touches on a lot of these points because the issues really do overlap. But overall, security, privacy, surveillance, and control are really big.

And then regulation and governance, this can go one way or the other. We need to have more regulation. More governance around these emerging technologies because of how important they are and how risky they are.

But at the same time, we need to not put so many regulations out that we don't develop these technologies at all. We've seen this with previous tech. Blockchain is one that NIST admitted that maybe they put regulations in place too soon and it caused it to falter a little bit.

Conclusion

Mehak: In the past, you usually had one emerging tech that was being created. There was time for organizations and governments to prepare.

But today's pace is a lot faster. Not only do you have just one emerging tech, you have two or three simultaneously happening, and the pace is so much faster. It really compounds the impact on society, on finance, on just about everything.

Keeping up is not just good, it is a requirement at this point for your organizations to succeed and to move into a post-quantum, post-AI future. AI is getting smarter. Quantum computing is coming. The basic question we have at the end of this is, are you and your organizations ready for what's coming?

Feel free to connect with us on LinkedIn. Thank you for joining and we look forward to connecting with you.

---

Answering Your Follow-Up Questions

Can quantum computers run without being at 0 kelvin? If not, how feasible is it to run them at 0 kelvin on this planet?

For now, the quantum computers we have typically operate at temperatures close to absolute zero in the range of 10 millikelvin. The primary reason for this is that qubits need to be isolated from all types of “noise” that can cause them to decohere, including, but not limited to, heat. Heat introduces energy into the system, which disrupts the delicate quantum states, leading to errors in computation. There are those in the quantum ecosystem working today to figure out how we can run quantum computers at higher temperatures while maintaining qubit integrity and functionality.

How can the integration of 6G technology and quantum computers revolutionize data transmission and processing? What are the potential challenges and opportunities in achieving this synergy?

The synergy between 6G technology and quantum computing promises a paradigm shift in secure, ultra-fast data transmission and powerful processing capabilities, paving the way for new applications and industries. Overcoming challenges related to technological maturity, infrastructure, standards, and security will be critical for realizing this vision.

In your view, how do quantum transformers impact the environment? Does the technology create more efficient machines that could lighten the energy load and energy consumption?

The overall environmental impact will depend on advancements in making quantum machines more energy-efficient and scalable. Innovations like room-temperature qubits or quantum systems with lower cooling requirements could shift the balance toward more sustainable computing.

Additionally, since quantum computers are expected to solve specific problems more efficiently, optimize supply chains, and enhance machine learning, amongst other optimizations, there is an opportunity to use less energy overall. This would be done by using less computational power, fewer emissions from vehicles used in supply chains, and faster cleaning of data for AI. Like with all of the emerging technologies discussed in this presentation, more research and funding (i.e., government and private sector) of research related to the minimization of the environmental impacts of these emerging technologies is needed.

Which markets are most likely to adopt quantum machines? Will this adoption resemble the current cloud service provider model (such as SaaS or QaaS) or will individuals have access to mini quantum computers within the next few years?

Quantum computing adoption will focus on industries like finance, banking, big tech, pharmaceuticals, defense, and logistics, where complex problem-solving and optimization are crucial. Cloud-based Quantum-as-a-Service (QaaS) will be the main avenue due to the high costs and technical demands of quantum computing hardware. Personal or mini quantum computers are unlikely to emerge in the near future, but are not an impossibility. In the beginning, most users will access quantum technology via cloud platforms like IBMQ and Google’s QaaS. In fact, you can already do this with IBMQ if you are proficient in quantum programming. If you are and haven’t coded in IBMQ yet, we’d highly suggest you try it!

You mentioned asymmetric encryption. Any thoughts on the reports of China cracking 22-bit AES a few days ago?

I love addressing this question and I also addressed it in the November 2023 CSA Quantum-Safe Security Working Group Crypto Newsletter. To start by giving credit where credit is due, they did take an incremental step in progressing quantum computing forward by utilizing a quantum annealer which isn’t often capable of executing an algorithm such as this.

However, the claims made of cracking encryption are not valid as they factored a small integer of 50 bits (not 22) which is a far cry from the 2048 bit keys and larger that are utilized today. Additionally, scaling from factoring 50 bits to 2048 bits is not a problem that’s 40 times harder – it’s exponentially more difficult. So, we can agree that progress was made, but the sensationalized headlines surrounding the progress made were not and are not valid.

What is the current state of quantum computing? How feasible is it for these computers to come to the mainstream, and by when? When will we have quantum personal computers?

There is a lot of discussion within the quantum computing ecosystem about when we will see a quantum advantage. Per the presentation, a quantum advantage is when a quantum computer will be able to solve a problem a classical computer cannot or cannot within a feasible amount of time. CSA's Quantum-Safe Security Working Group, where Mehak is a co-Chair, estimates quantum computers showing a quantum advantage in 5-6 years.

Quantum computers are already available via cloud environments. For example, if you are familiar with quantum coding, you’ve been able to access IBM’s quantum computer for years. In terms of mainstream use, it will start with large organizations and governments with a trickle-down effect for common consumers further down the line once a larger quantum computing infrastructure can be created.

Creation of this infrastructure is dependent upon several factors, including which quantum computing technology shows a quantum advantage first and which quantum computing technology the larger players are willing to develop on a larger scale. This could well be two or more different quantum computing technologies.

As for personal quantum computers, I see this becoming more viable if we are able to figure out how to stabilize quantum computers outside of nearly 0 K temperatures and vacuums. This technology is being researched and with more funding may have the potential to develop faster. I would keep an eye out for a quantum advantage, after which, similar to the Artificial Intelligence boom of 2023, I expect we will likely see a quantum boom.

How can companies develop a robust process for implementing quantum encryption methods to stay ahead of potential vulnerabilities?

Companies can adopt post-quantum cryptography, assess risks, use flexible encryption frameworks, automate updates, collaborate with researchers, and invest in quantum-safe infrastructure. CSA has a Quantum-Safe Security Working Group that has guidance on preparing for quantum computers being readily available. You can also review this article I, Mehak, wrote a few months ago about the finalization of NIST FIPS 203, 204, and 205, which are the 3 Post-Quantum Cryptography algorithms suggested by NIST to be utilized by organizations. The article also includes helpful links to start your post-quantum journey.