Quantinuum, the world’s largest quantum computing company, and HSBC announced a series of exploratory projects that exploit the potential near- and long-term benefits of quantum computing for banking with specific projects in cybersecurity, fraud detection, and natural language processing. According to a story in MarketsMedia news, an initial exploration of the multi-stage collaboration is intended to demonstrate the use of quantum computing-hardened cryptographic keys, including uniquely combining them with post-quantum cryptographic algorithms to mitigate current and future cyber threats. This effort to strengthen resiliency against advanced cyber threats is increasingly critical as the transition point between the capabilities of classical and quantum computers continues to approach. Quantinuum’s Quantum Origin is a platform that uses the operations of a quantum computer to strengthen the cryptographic keys that are used to protect transactions and identification processes. Quantum Origin is deployed on existing “classical” cybersecurity infrastructure and is the first commercial product on the market that uses a quantum computer to produce provably unpredictable cryptographic keys, which could offer an extra layer of security to protect HSBC’s most valuable data. HSBC and Quantinuum will run Quantum Origin via an HSM provider. In a second part of the collaboration, HSBC and Quantinuum will research and explore the potential benefits of quantum machine learning (QML) and quantum natural language processing (QNLP) for HSBC’s business. With fraud detection as an HSBC priority, the collaboration will examine advanced QML techniques that are enhanced by qubit routing and circuit optimization techniques provided by Quantinuum’s architecture-independent software development platform, TKET. Additionally, HSBC and Quantinuum will explore QNLP, which is a novel form of language-based AI that uses an explainable model, rather than the “black box” methods of traditional classical large-language models. QNLP will be based on training quantum states and processes that encode word meanings. This approach may enable native NLP tasks such as question answering or text similarity, which could be valuable in regulated markets dealing with customer data.