This post has been contributed by Dr Jennifer Dickfos, a Senior Lecturer in the Department of Accounting, Finance and Economics and a member of the Law Futures Centre.

AI’s impact on the Australian insolvency profession to date has been limited.  This is evidenced by three surveys of insolvency practitioners from 2017 to 2020.

The results of a national online survey  of registered bankruptcy trustees and registered liquidators in Australia during July 2017 to February 2018 were published in AI and the Insolvency Profession: The State of Play (2018) 26 Insolv LJ 172 . Despite the existence of “Reg Tech” and “Standardised Business Reporting”, progress towards a digital insolvency practice was slow. Recommendations made at the time, in order for the development of digital practice to be successful were for further education of insolvency practitioners regarding AI’s potential; innovation to be practice directed; and stakeholder collaboration.

The lack of progress in Australia was consistent with what was not happening overseas.

The International Association of Restructuring, Insolvency & Bankruptcy Professionals (INSOL International) carried out a member survey in 2018 -2019 to determine the degree of technology use in both global bankruptcy and restructuring practices.  Survey results were reported in The Role of Artificial Intelligence (AI) and Technology in Global Bankruptcy and Restructuring Practices which confirmed that the European and United Kingdom insolvency professions were similarly not proactive in adopting technology. A shared perception with the Australian survey was the need for collaboration between lawyers and accountants to adopt technology.

The latest survey results of the 2020 Macquarie Insolvency industry pulse check (which surveyed 112 insolvency  firms predominantly in New South Wales, the Australian Capital Territory and Victoria ) confirms that the majority of insolvency firms were not focused on automation and technology as current challenges.  Their technology use was generally restricted to automation of previous manual activities such as document management, scanning tools (62%) and workflow tools, with fewer than 23% of firms using AI such as data analytics or predictive systems to improve the client experience and make more informed strategic decisions.

Reasons provided in the 2020 Macquarie Insolvency industry pulse check for the limited technology adoption to date were: industry conservatism, hesitation to change, lower recent revenues and no evidence of industry peer leadership.  The survey authors’ recommendation was to look to other professional sectors integrating data analytics and predictive systems into their daily processes. 

In looking at other professional sectors such as lawyers, there is a perception however, that not all legal tasks can be automated. In their empirical study, Can Robots Be Lawyers? Computers, Lawyers, and the Practice of Law, Dana Remy and Frank Levy propose that for a computer to automate a lawyer’s tasks, such tasks must be structured or routine. To replicate more advanced decision-making, statistical models using algorithms are created. To strengthen their predictability, these algorithms are refined through a process of “machine learning” although their predictability is limited as Remus and Levy note that such statistical models “have difficulty processing contingencies that lie outside the data on which they were trained”. Based on their empirical study Remi and Levy identified the following lawyers’ tasks as capable of automation: document review and drafting, due diligence, legal research, analysis and strategy.  On the other hand, they suggested that lawyers’ tasks such as case management, fact investigation, advising clients, court appearances and negotiation which require unstructured human interaction and/or emotional intelligence cannot currently be performed by computers.

Applying Remus and Levy’s model to insolvency practices the following tasks are considered capable of automation: real and personal property searches in bankruptcy and liquidations; calculation and collection of income contributions in personal insolvency; and creation of a bankruptcy smart form which would automatically populate sections of a creditors’ report. Tasks such as negotiating with creditors or advising on the potential restructuring of a failing business are not considered capable of being automated at this time.

However, those tasks which are capable of automation provide opportunities for technology to increase productivity and lower costs in insolvency administrations. Such benefits should be pursued vigorously, especially when the cost of insolvency administrations is a common complaint, regularly levelled at the insolvency profession.