After decades of theorising, quantum computers are finally emerging. In December, it was announced that a team from the University of Science and Technology of China had built a quantum computer capable of performing computations nearly 100 trillion times faster than the world’s most powerful supercomputer.
In October 2019, Google announced it had developed a quantum device that took 200 seconds to sample data that it claimed would have taken a standard ‘classical’ supercomputer 10,000 years. Meanwhile, IBM has been operating quantum computers since 2016; Microsoft has eight quantum laboratories (in Europe, the US and Australia); and Canada/US-based D-Wave offers access to its quantum computers via the cloud.
So how does this matter to accountants? Troels Steenstrup Jensen, head of machine learning and quantum technologies advisory at KPMG Denmark, points out that auditors can leverage quantum computing power to analyse data in ways that standard computers find too difficult.
Take ‘round-tripping’ transactions, where one company inflates its sales figures by selling assets to another and subsequently buying them back, often via a chain of intermediaries, for the same price. Similarly, complex transactions can be used to hide profits or costs. Detecting such wrongdoing involves a ‘hard maths problem, but you can do it with quantum computing’, says Jensen.
‘One of the biggest challenges faced by auditors is spotting accounting fraud in isolation without any insight from other firms’
Shungo Miyabe, senior manager for data and AI at KPMG, says quantum computers can help auditors by analysing data in graphic form and drawing insights from their illustration of relationships between products and financial transactions, especially where more than two related companies are involved.
‘This allows us to discover structural properties in data that are not evident using traditional data analytics methods,’ Miyabe says. ‘Advances in graph-learning algorithms show great promise for detecting accounting fraud.’
Because graphics can be built from encrypted data, they can include shared information from different entities without breaking the confidentiality and privacy regulations that apply to raw financial data.
This could be critical. ‘One of the biggest challenges faced by auditors is spotting accounting fraud in isolation without any insight from other firms,’ Miyabe says.
Once inside an IT system, quantum-assisted hackers can break encryption that protects financial transaction data
However, the quantum technology revolution brings threats as well as opportunities.
Quantum computers can be used by malicious actors as well as legitimate businesses. Once inside an IT system, quantum-assisted hackers can break encryption that protects financial transaction data. This is as much a threat to national security as it is to sensitive company data.
A Deloitte study warns of the risks to security and advises organisations – including governments – to begin planning for hacks now and creating quantum teams, including the potential and risk of quantum computers in strategic planning meetings.
Narayanan Vaidyanathan, ACCA’s head of business insights, agrees that quantum hacks are real risks, but points out that accounting practices and their clients have ‘a certain amount of runway time to prepare’.
‘Now is a good time for accountants to acquaint themselves with quantum computing and how it could pose a risk to the security of data held by practices and clients.’
Why are quantum computers so powerful?
Quantum computers deal in qubits – the equivalent of bits in classical computing.
Unlike a bit, a qubit can hold more than one mathematical value at the same time (the ‘superposition’ property in quantum physics), allowing a quantum computer to process a vast number of potential outcomes simultaneously rather than sequentially.
Combined with the ability of linked qubits to share a single state, so that a change in one instantaneously changes the state of the others in a predictable way (‘entanglement’), this means that quantum computing power increases exponentially with the number of qubits in use, rather than linearly as in bit-based classical computing.