AI and the future of law

Professor Kimberlee Weatherall, Associate Dean, Research

Professor Kimberlee Weatherall says the trademark system might not seem like an obvious starting point for a discussion of the impact of AI on the law or government decision-making. The risks are not a matter of life and death. They are contained within economic and innovation policy. But she says it’s preciously for this reason – “because it is mundane” – that the trademark system is an early candidate for automation, with datasets and technologies (such as content-based image recognition) already moving in that direction.

In her latest research, Professor Weatherall unpacks two alternative approaches to implementing automated decision-making in the trademark system. One approach, she explains, involves trying to make an intelligent/autonomous system do what humans used to do. A second and far better approach, says Professor Weatherall, is to imagine how the trademark system could be reshaped to better serve its purposes by using the capacities of new technologies.

“The goal here is mutual accommodation,” she says. “Simple automation of existing legal processes will not work, because the legal rules are not computable, no matter how good the technology.”

“But that does not mean we should simply throw up our hands and say, ‘computers can’t do it’,” she continued. "Instead, we should look for ways to adjust our legal systems.”

“This is going to require some shift in the law to use technical capacity, as well as some humility on the part of those designing technology to avoid trying to take-over the entire legal system.”

Professor Simon Rice, Director, Professional and Community Engagement

In recent years, AI has been presented as a solution for people without access to a lawyer. While previous inclusion methods have encompassed things like drop-in sessions, telephone advice lines, pamphlets and brochures, DIY kits, video links and websites, AI has the ability to offer much more sophisticated and personal help through ‘access to justice’ mobile applications.

But Professor Simon Rice says the solution isn’t so simple. AI may be inaccessible for the same reasons that access to justice is inaccessible, such as cost, language, age, culture and disability, he explains.

“In the absence of established standards, there is no certainty that an ‘access to justice’ app will itself be accessible,” said Professor Rice.

“A further consideration is the growing awareness of bias in the data on which AI’s algorithms work, and the resulting discriminatory effect,” he continued. “This effect could be especially acute when the users of an app are from the same marginalised and disadvantaged communities that are not represented in the data."

Professor Rice says that human-centred design requires consultation with user communities, and design by and for them. Without careful consideration, we threaten to thwart the effectiveness of AI for access to justice, he said.

Dr Carolyn McKay, Lecturer in Law, Civil & Criminal Procedure and Criminal Law

Judicial evaluations are increasingly complemented by AI tools to provide accurate predictive capabilities and assessments of risk, for instance, in bail, sentencing, rehabilitation and parole procedures.

“Assessments have traditionally been the function of human discretion and the intuition of judicial officers, based on clinical assessments, framed by legislation and common law principles, and encapsulating the concept of individualised justice,” explained Dr Carolyn McKay.

She says that while there is a recognised need for responsible, accountable and ethical algorithmic design and instruments, there is also considerable risk when it comes to matters of individual liberty, justice and public safety.

“The use of AI can assist in determining the extent of liberty granted in criminal proceedings,” said Dr McKay. “But the proprietary nature of AI devices means the calculation of the risk score is not very transparent, and unknowable to both the offender and the court.”

Dr McKay says the Geneva conference provided a great opportunity to comprehensively focus on AI from multiple perspectives and world views.

“It’s an evolving area of criminal procedure, and we need to ensure that we understand all of the issues at play to ensure accountability and fair and accurate outcomes,” she explained.

Dr Penelope Crossley, Senior Lecturer and Member, China Studies Centre

The digitisation of the energy market has led to a fundamental transformation currently underway in the sector globally.

‘Energy 4.0’, derived from the Fourth Industrial Revolution, is changing the traditional roles of energy market regulators, market participants and end-consumers. Characterised by cyber-physical systems, such as smart grids, and the new technologies and structures – including AI, big data, blockchain and smart contracts – the revolution has left the industry in disruption.

Dr Penelope Crossley is researching the risks, as well as priorities, for energy market regulation in an era of rapid technological development. 

“The challenge is three-fold: protecting the consumer and supporting innovation, while understanding the potential risks and consequences associated with disruptive technologies,” said Dr Crossley.

“We need to address the complexities of these emerging technologies, with the changing role played by consumer protection, to ensure the continued development of competitively functioning energy markets.”