Using their “Quantum NLP” or “QNLP” framework, once the meanings of words and phrases were encoded as quantum states and processes, they were able to prepare quantum states that encode the meaning of grammatical sentences on quantum hardware. Posing a question to the quantum computer, constructed by the vocabulary and grammar the quantum computer has learned, it returned the answer. In subsequent work, Coecke and colleagues developed ways to implement such QNLP models on today’s NISQ devices.
This compositional way of endowing grammatical structure to the quantum model may help to validate and verify the model’s performance by giving the model the means by which to be interpretable. It will also provide insight into the inner workings of the model in regulated sectors such as finance, legal, and medicine, where transparency is critical. By contrast, classical algorithms in NLP and the broader field of machine learning have raised concerns due to their “black box” nature. In legal terms, it is hard to understand how a machine learning system has come to a certain prediction or decision. Coecke and his team argue that quantum computers can enhance NLP thanks to discovering that language is “quantum native.” He suggests this will move the world away from the opaque “brute force” techniques at the heart of current NLP.
The legal applications of NLP are many, and the advancement that Coecke and others are making will have profound impacts on the legal field. They may, for example, increase the average person’s access to the legal system, thus helping to close the access to justice gap. QNLP may also help to reduce the incidence of contract disputes. Vagueness in language is a significant cause of unintended interpretation. In today’s increasingly networked society, contracts need to be understood by parties operating in different languages around the globe. Predictive NLP models for court decisions were 79% accurate in 2016 and are becoming increasingly accurate.
Likely the most common use for NLP in the legal field has been document review. In corporate litigation, millions of documents may need to be reviewed for relevance. E-discovery systems collect, analyze, and store millions of documents for pre-trial discovery. These documents can range from memos and contracts, blueprints, and CAD drawings, to instant messages and email. As a result, they are challenging to categorize. QNLP will allow for streamlining legal research, helping litigators develop their cases earlier thanks to its potential for speed, accuracy, and intuitiveness. Further, QNLP will help draft and analyze legal documents, automate routine tasks, and predict rulings. While online legal databases have been available for decades, QNLP will help streamline the research process and anticipate the most relevant findings to a favorable ruling. There is already an app “DoNotPay” that uses NLP to automate the litigation and defense process for a variety of scenarios.
Cybersecurity
The Importance of Cybersecurity
As information storage and analysis have moved increasingly online, cybersecurity has moved to the forefront of legal consideration. Privacy violation or financial loss through hacking prompts assignment of guilt including breach of contract and negligence. The implications for security are then immediate: Gaining control of a computer system can mean compromised critical infrastructure, or damage to financial, or healthcare systems. As cybersecurity (and hacking) technology develops, so too must the legal system.
How Will Quantum Computing Affect Cybersecurity?
There is a close relationship between quantum computing and cybersecurity. One of the first quantum algorithms invented, Shor’s algorithm, factors numbers into their constituent primes. This is significant because the RSA algorithm—the basis for much of the internet’s cybersecurity—uses such factoring as its mathematical foundation. Conventional computers require vast time and resources to factor very large numbers. It is straightforward to use numbers that are sufficiently large that there is no realistic prospect of building a classical computer that can factor them. On the other hand, quantum computers running Shor’s algorithm gain power exponentially with the addition of qubits. At some point, a quantum computer of sufficient power will run Shor’s algorithm and undermine the foundations of RSA encryption. This is already the subject of much attention, and alternative algorithms are already being developed. Nevertheless, there is a very real possibility of “quantum hacking” in the coming years, and indeed it is already accepted that encrypted material sent over the internet today can be intercepted and decrypted later using a quantum computer—a so-called hack now, decrypt later attack. The ability to break public-key encryption and thus expose the secrets of every person, corporation, and nation-state is legally alarming. This risks impacting everything from military communications to bank accounts to private emails.
Confidential information must be protected from “bad players” such as criminals or terrorists. This is particularly true when the stakes are very high, such as when protecting the privacy of a whistleblower or witness. Fortunately, there are already new “post-quantum” cryptographic algorithms being developed that are believed to be invulnerable to both classical and quantum-based attacks. These algorithms are being studied as part of a multiyear NIST program to find quantum-safe algorithms optimized for different use-cases. NIST is currently considering seven finalist algorithms. It is expected to announce at least one winner in the next few years.
There are also quantum communication technologies being developed that offer stronger security guarantees than their classical equivalents. For instance, “quantum key distribution” networks transmit quantum data and exploit the laws of quantum mechanics that make undetected eavesdropping and tampering impossible. These networks are already being developed using fiber optics or satellites and lasers, including the Chinese “Micius” satellite launched in 2016. In parallel work, quantum-generated encryption keys are now commercially available that support standard and post-quantum algorithms. These keys are superior to their classical counterparts because they are generated from a verifiable quantum source, which makes them provably unpredictable.
Such advances in communications security raise legal questions that do not have an obvious answer. Governments may find they encourage adoption of such standards for others, while continuing to grapple with the desire to uncover threats and to eavesdrop on adversaries. It may turn out, in the long run, the laws of physics prevent such security activity.
In addition to quantum cybersecurity technology itself, investment in the sector has also become a legal focus. Foreign direct investment (FDI) and cross-border M&A are coming under increasing scrutiny worldwide, particularly where sensitive or “emerging” technologies are concerned. Foreign investors and companies focused on the quantum computing sector should ensure they are well-advised on these rules sufficiently prior to entering into discussions to raise foreign capital in fundraising or acquire or sell to foreign investors.
Conclusion
Quantum computing will have a significant impact in the coming years. It will affect the whole of scientific and technological progress, as well as many of the trappings of the modern economy and society as a whole. The legal aspects of intellectual property, natural language processing, cybersecurity, national defense, and others will each need to evolve to match the technical development. Lawyers interested in this area should contact technical experts to develop policies rooted in sound scientific basis.
