'Robot Lawyers': How Far Will Automation Go?

1 May 2022

It is no understatement to say that artificial Intelligence (AI) can transform legal services. With AI comes the promise of greater efficiency in that it can be used to automate legal tasks in a way that creates faster and more consistent outputs than human lawyers. This has the potential to make legal services significantly cheaper and more accessible to customers, thereby making justice accessible more broadly. 

However, AI has also raised significant social questions. If AI can displace human workers, then what will be left for humans to do? A study conducted by McKinsey in 2017 estimated that between almost 0% to 30% of the total hours worked globally could be automated by 2030 depending on the speed in which AI is adopted in industry. This is a rather startling figure. Nonetheless, the conclusion is undeniable: The impact of AI on the world of employment will be significant and can lead to widespread unemployment if not handled carefully.

The legal profession is not immune from the potential job-displacing effects of AI. In fact, McKinsey has estimated that that 22% of a lawyer's job can be automated. The figure is 35% for law clerks. So, the question must be asked: How far will automation go? How widespread will job displacement become? Will this be to the extent where we see the rise of 'AI judges' deciding cases in courts?

Although AI has raised significant challenges, and as despondent as the statistics may seem, the future may not be as gloomy as what we think. And here's why.

Firstly, let's examine the work that lawyers actually do. Here is one key aspect of the debate:

Not all legal work is the same. 

Legal work ranges in complexity. On one hand, lawyers may be engaged in 'mechanical' legal tasks such as document review and contract drafting. On the other hand, lawyers may be involved in testing the fringes of the law in bringing novel cases in courts. In terms of middle ground, a lawyer's job may involve both of these things: A combination of mechanical tasks in addition to working on more exciting projects with less clear-cut answers. So where does this leave AI? 

The first type of work lends itself to a greater risk of automation. Consider contract review as an example: In 2017, JPMorgan announced it had used contract intelligence software which could perform document review tasks in seconds even though it typically took human lawyers 360,000 hours. Tasks such as these can be automated because AI can be trained on millions of existing documents, contracts, case files and legal briefs to flag appropriate issues that may arise in the contract.

Source: Pixabay (2022)

However, automation is less likely in areas where there are still a large number of grey areas in the law or where the fringes of the law are being tested in litigation. Lawyers may in fact be using existing legal causes of action to seek redress for harm with judges needing to rule on such matters. A rather prominent example in the media has the recent Sharma litigation in the Full Federal Court, with the applicants testing the frontiers of the existing tort of negligence to bring a claim against the Federal Environment Minister for future climate harm. In such cases, there isn't necessarily a correct answer. As such, this type of legal work requires ingenuity and creative thinking, with lawyers extrapolating on incomplete information to extend the boundaries of existing legal frameworks beyond what is currently accepted law. Such circumstances are examples of where AI cannot be applied in a clear-cut manner to automate tasks. 

In sum, when one says that 'AI will automate law', a little more nuance is needed in the debate. While AI can automate more mundane legal tasks on an outright basis, there are types of legal work that cannot currently be automated. The debate should not be focused on AI completely taking over human lawyers in the legal profession. Rather, AI should be viewed as a complement to human labour, whereby AI can automate work that is horrendously labour-intensive, repetitive and low-skilled in nature. In essence, AI frees lawyers to pursue more creative and constructive tasks. 

Now, this is not to say that jobs won't be impacted. One only needs to examine the issue with the eyes of a law graduate to see the point. Since law graduates are typically introduced to legal work through 'low-skilled' tasks such as due diligence and document reviews, where does automating these low-skilled tasks leave law graduates or other legal practitioners with a more junior status in the profession? Is it now going to become even harder for graduates to get their foot in the door? 

And here comes another key point: 

The nature of work will change.

Continuing on the theme of law graduates, there is no reason why they cannot be trained in using new AI systems. In fact, it has even been suggested that AI can help make entry-level lawyers better practitioners faster. AI-based legal research assistants such as CaseIQ can provide iterative feedback on a brief by suggesting changes and providing additional documents for the lawyer to consider when formulating their arguments. Further, freeing time may allow law graduates to progress onto more complex tasks faster, thereby improving their rate of learning.

Now, what about other lawyers apart from law graduates? Well, now that lawyers have more time on their hands, there may be scope for them to service new clients they previously may not have had the time to serve (after all, there is no shortage of legal grievances!!). Further, the rise of AI and technology carries demand for new skillsets and therefore new jobs. Influential scholar Richard Susskind has predicted that new types of legal professionals will replace conventional lawyers. He provides the example of 'legal technologists' who are self-sufficient in their understanding of technological matters and can readily apply this to legal contexts. Furthermore, he also provides the example of 'legal engineers' who will design internal systems and standard working practices to conform with the law. Thus, it can be said that while 'conventional lawyers' may start to decline, lawyers will take on new roles in the future. 

Final Thoughts

The AI debate needs careful, nuanced consideration. Saying that legal tasks will be automated is a completely different thing from saying that human lawyers will become redundant and that justice will be purely AI-driven. While it is probable that AI will replace human labour in fields that require more mundane, repetitive tasks, AI has the potential to free up significant resources and allow lawyers to pursue more creative tasks. And all of this is not to mention that the rise of AI will create demand for new skillsets and therefore new opportunities. 

When having a conversation about AI, we need to be cool, calm and collected. And above all, we need to be measured. 


[1] Erin Winick, 'Lawyer-Bots are Shaking Up Jobs', MIT Technology Review (online, 12 December 2017) <https://www.technologyreview.com/2017/12/12/105002/lawyer-bots-are-shaking-up-jobs/>.

[2] LexisNexis, 'Lawyer vs AI: A Legal Revolution' (online, 19 April 2018) <https://www.lexisnexis.com.au/en/insights-and-analysis/practice-intelligence/2018/Lawyer-vs-AI-A-legal-revolution>.

[3] Louise Camenzuli, Julia Green & Max Newman, 'Sharma Appeal Decision: End of the Road for Novel Duty of Care?', Corrs Chambers Westgarth (online, 16 March 2022) <https://www.corrs.com.au/insights/sharma-appeal-decision-end-of-the-road-for-novel-duty-of-care>.

[4] James Manyika et al., 'Jobs Lost, Jobs Gained: What the Future of Work Will Mean for Jobs, Skills and Wages', McKinsey & Company (online, 28 November 2017) <https://www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-of-work-will-mean-for-jobs-skills-and-wages>.

[5] Elizabeth C. Tippett & Charlotte Alexander, 'Robots are Coming for the Lawyers - Which May be Bad for Tomorrow's Attorneys but Great for Anyone in Need of Cheap Legal Assistance', The Conversation (online 10 August 2021) <https://theconversation.com/robots-are-coming-for-the-lawyers-which-may-be-bad-for-tomorrows-attorneys-but-great-for-anyone-in-need-of-cheap-legal-assistance-157574>.

[6] Michael Legg & Felicity Bell, 'Artificial Intelligence and the Legal Profession: Becoming the AI-Enhanced Lawyer' (2019) 38(2) University of Tasmania Law Review 34. 

[7] Richard Susskind, Tomorrow's Lawyers: An Introduction to Your Future (Oxford University Press, 2nd ed, 2017). 

Cover image sourced from here.

Write & Read to Earn with BULB

Learn More


It is an interesting article. Managing AI ethically would help automate the legal process, cut corruption and wrong judgments. AI could solve many crime investigations because it can process large amounts of data quickly.