What AI means for the future of quantum
In 2014, when the UK launched its National Quantum Technologies Programme, the government was taking a bet on a “distant” technology with great transformative potential. Since the launch of the programme, the UK has invested over £1bn and continues to invest in and support the development of quantum technologies, spanning computing, sensing, and communications. In that time, we have also seen the exponential rise of AI, now becoming a technology that is increasingly commonplace. As both technologies mature, they are growing closer together. The excitement around their convergence is understandable, but it has not always been matched by clarity about what is near-term and what remains a longer-term prospect. To seize the opportunity presented by the convergence of these two technologies, we need to understand the existing gaps and how we address them.
AI for quantum computing
The relationship between quantum computing and AI runs in two directions, each on very different timelines, with the less discussed of the two delivering the more near-term benefits. AI is already being applied to improve quantum hardware by correcting errors, optimising circuits, and supporting hardware design. Quantum systems are inherently noisy, and machine learning is one of the more promising routes to managing that. Progress here is real but tends to be overshadowed by the bigger headline claims.
More of the excitement, and more of the hype, surrounds the reverse direction: using quantum computers to enhance AI. The potential in areas such as large-scale optimisation and machine learning exists, but the distance between where quantum hardware currently sits and what would be required to make it a reality is quite large. Without fault-tolerant quantum computers, which the field is still far from achieving, meaningful quantum advantages for AI are not near-term prospects, and should not be the only focus when considering the potential of AI and quantum.
AI for quantum sensing
Much of the quantum-AI conversation centres on computing, but sensing tells a different story. Quantum sensing does not depend on fault-tolerant hardware, which means it is already generating real-world results. Quantum sensors can detect tiny variations in gravitational fields with enough sensitivity to map underground pipes and infrastructure without breaking the surface. They can measure the magnetic activity of the brain in real time. They can support accurate navigation and timing that does not rely on GPS. In each case, what they produce is data that was simply not available before.
More data, however, is only useful if you can make sense of it. Processing the high-resolution and high-volume output of quantum sensors is not straightforward, and this is where AI plays a meaningful near-term role by interpreting what the quantum sensors produce. Together, the two technologies offer something neither can deliver alone: hardware that measures the world with unprecedented sensitivity, and software that turns those measurements into actionable.
How we capitalise on the opportunity
The UK has built real capability in both quantum and AI, but largely along separate tracks. The Compute Roadmap, the National Quantum Strategy and the Industrial Strategy have each been integral to progress but have limited focus on convergence. The gap is less about something being missing and more about coordination between what already exists. It also helps to keep "AI for quantum" and "quantum for AI" clearly apart, because they are at very different stages and raise very different questions for both industry and government.
Standards are the most practical place to start, and the area where the UK has the strongest claim to lead. The UK has a long standards heritage and role in bodies like BSI and the Internet Engineering Task Force (IETF) , making convergence exactly the sort of area where being early shapes how the market will eventually operate. Standards would also solve a problem the sector is already running into. Claims about AI in quantum systems are difficult to compare, commercial roadmaps are often kept quiet, and buyers have little basis on which to judge one product against another, particularly where AI is interpreting sensor output that has never been measured before. There is a related question about quantum data, which may need new classifications and clearer rules on access before anyone can build usable datasets from it. The caution is against fixing technical detail too early, which is why faster and more provisional routes, such as a Publicly Available Specification, are worth looking at.
As with many emerging and transformative technologies, the UK needs to get the skills, investment, and infrastructure right to make the most of the promise of quantum and AI together. Capitalising on this opportunity will require moving away from the hype and following what the evidence is already demonstrating. Quantum sensing, supported by AI, offers something more tangible: not a bet on a future breakthrough, but a return on progress that is already underway. The question for the UK is whether it is directing enough attention towards the opportunity that is in front of it now.
If you would like to contribute to techUK's policy position on the future of AI and quantum, please contact sara.duodu@techuk.org.
