The head of the Australian Law Reform Commission (ALRC) has challenged the narrow parameters of the debate over AI regulation, warning of a focus on form over substance in the regulation of the emerging technology and the danger of being smart, rather than wise.
Federal Court Judge and president of the ALRC Mordy Bromberg, in a speech to the Australian Law Forum, noted that technological change was often a driver of legal reform, pointing to the ALRC’s current work on overhauling human tissue laws, including a possible change in how death is defined. But AI — “a different species of technology” — promised something on a much greater scale:
Artificial intelligence has the capacity to affect and to influence human behaviour on a scale which we have never before experienced. Because every law in one way or another seeks to regulate human behaviour, an instrument with that capacity will likely touch most of our laws.
Such an instrument comes with significant risk, given “the likely ubiquity of AI and its forthcoming role as the dominant purchasing and transactional agent for us all, together with its capacity to become the primary source of our information and news”. Scoping out the risks, identifying who bears responsibility for them, the unique properties of AI and determining whether we should take a preventive or remedial approach to damages — rather than vaguer questions like whether to take a “light touch” approach — amounts to “the largest and most complex law reform exercise that currently confronts Australia”.
Bromberg is hardly a legal Luddite. “AI is and will remain an agent of change that will transform the legal profession. And so it should … It may be that AI will result in fewer lawyers. We ought to be enthused by those features of AI that will enhance the services we provide. We are, however, somewhat threatened by the notion that AI will replace us … I am personally optimistic that AI can and will make very substantial contributions to the productivity of legal professionals.”
Related Article Block Placeholder
Article ID: 1217502
But he distinguishes between the intelligence that AI displays and the wisdom that complex legal cases demand. “Wisdom is a skill of a greater degree of order than mere intelligence. It is far harder to be wise than it is to be smart. And wisdom is a skill developed and exercised not merely by reference to data but by reference to innately human experiences that AI does not share.”
Independent. Irreverent. In your inbox
Get the headlines they don’t want you to read. Sign up to Crikey’s free newsletters for fearless reporting, sharp analysis, and a touch of chaos
By continuing, you agree to our Terms & Conditions and Privacy Policy.
Bromberg’s distinction is an apt one for the quality of debate around AI regulation, which at the moment is dominated by economic perspectives: the AI industry itself — spruiking wildly inflated claims of productivity benefits to be achieved by handing them a lot of public money and not regulating them, trade unions trying to hobble AI so no worker ever loses their job, and spruikers like the Productivity Commission, who lean toward minimal regulation regardless of the impacts on creators whose hard work has formed the basis of the profits of AI companies.
That economic focus seems akin to the regulators of 2008 focusing only on the potential advertising revenue impacts of social media — rather than the potential for disinformation, online bullying, the risks posed by social media usage to mental health or its ability to foster political division. The mental health risks of AI are already becoming apparent — what happens when articles like these go from the stuff that makes you slap your forehead at how dumb people are, to describing a wide-scale mental health issue, as we’ve already seen with the impact of social media on young people? AI is also developing rapidly at the same time as we are still trying to play catch-up in other key areas where politicians have failed to implement previous law reform challenges, like privacy, which will be dramatically affected by AI.
As with social media in 2008, the benefits and damage from AI can only be guessed at. We know from the history of new media that it’s difficult to predict how it will shape society — and also that people will make huge claims about both the positives and negatives of the new medium for self-interested or ideological reasons. AI, as a new kind of technology rather than merely a new medium, is that challenge on steroids.
But that makes Bromberg’s essential point — that we should be undertaking a basic stocktake of AI’s impacts and their legal implications — all the more powerful. If we wait, we might have a clearer idea of what we need to regulate, but we’ll also — as we’re doing with social media — be “chasing a tail that can never be caught”.
What will be the main challenges of regulating AI?
We want to hear from you. Write to us at letters@crikey.com.au to be published in Crikey. Please include your full name. We reserve the right to edit for length and clarity.
