Back
Image Alt

The Singapore Law Gazette

Three Big Ideas About AI for Lawyers

Introduction

In the late Professor Stephen Hawking’s final book, Brief Answers to the Big Questions, the eminent physicist provided his personal take on some of the biggest questions in science. On the question “Will artificial intelligence outsmart us?”, Professor Hawking opined that if computing power continued to increase at the same rate as indicated by Moore’s Law, it was likely that computers would “overtake humans in intelligence at some point in the next hundred years”.1 Stephen Hawking, Brief Answers to the Big Questions (London: John Murray (Publishers), 2018), at p 184 (“Brief Answers”).

If Professor Hawking’s prediction proves to be correct, it will take a few generations before AI outsmarts lawyers. But the short to medium-term impact of AI is likely to be significant in light of three big ideas about AI: AI as a utility, AI as augmentation, and AI as legal revolution. As the legal profession begins its journey into the brave new world of 2019 and beyond, it is worthwhile for lawyers to take a step back and reflect on how these three big ideas about AI will impact on the future of legal practice.

AI as a Utility

Amidst the AI hype, some commentators view AI as a kind of utility (like electricity) which will silently underpin the operations and products of businesses.2 See e.g. Thomas H. Davenport, The AI Advantage: How to Put the Artificial Intelligence Revolution to Work (Cambridge, Massachusetts: The MIT Press, 2018), at pp 6-7 (“The AI Advantage”). This “important but invisible” effect of AI will not result in many groundbreaking innovations, at least in the short run.3The AI Advantage, id, at pp 6-8. As Thomas H. Davenport observes in The AI Advantage, “[i]n the short term, AI will provide evolutionary benefits; in the long run, it is likely to be revolutionary”.4The AI Advantage, supra n 2, at p 7.

Still, AI’s potential ubiquity suggests that it will be difficult for lawyers to opt out of the AI conversation. Already, different types of AI technologies are being applied across a wide range of applications. For example, deep learning (a highly complex form of machine learning) is being used in image and voice recognition, while physical robots are used for factory and warehouse tasks.5The AI Advantage, supra n 2, at pp 11, 14 and 16.

Should such AI technologies become pervasive across all industries, it would be hard to imagine that AI issues would not arise in disputes or contracts, or that clients would refrain from asking their lawyers for legal advice on AI issues.

Reflection point: Lawyers should carefully consider how to improve their understanding of AI to meet their clients’ future needs (for a basic introduction to AI, see our September 2018 article titled “What is Your Artificial Intelligence Quotient?”).

AI as Augmentation

Since AI’s recent re-emergence, the perennial question for the legal profession is whether AI will ultimately replace lawyers. Not all commentators take the view that AI will only impact “the lower end of the skill and economic scales”, such as paralegals and junior lawyers.6 Nigel M. de S. Cameron, Will Robots Take Your Job?: A Plea for Consensus (Maiden, MA: Polity Press, 2017), at p 30. In time, sophisticated algorithms may also take charge of more nuanced tasks such as judgment and decision-making.7Id. Proponents of “large-scale automation” have taken the view that “AI was improving rapidly, [while] humans were not”.8The AI Advantage, supra n 2, at p 130.

However, things may not be as simple as they seem. The AI hype has largely been generated from AI’s success in beating human players at board games, but it is much more difficult to apply AI to mundane human activities such as “properly answering a question over the phone”.9 John Paul Mueller and Luca Massaron, Artificial Intelligence for Dummies (Hoboken, NJ: John Wiley & Sons, Inc., 2018), at p 218.

Even if “a job was technically automatable”, a 2017 McKinsey study showed that other factors affected whether automation would in fact occur, for example, the “costs to automate” and “the relative scarcity, skills, and cost of workers who might otherwise do the activity”.10The AI Advantage, supra n 2, at p 132.

Hence, some commentators favour large-scale augmentation instead, which involves human-machine collaboration. For example, Thomas H. Davenport argues in The AI Advantage that large-scale augmentation is more likely to occur than large-scale automation because “AI tends to support or automate tasks, not entire jobs”.11The AI Advantage, supra n 2, at p 133. In addition, while automation will render some aspects of the job redundant, “new roles and skills” or even “entirely new jobs” can be created.12The AI Advantage, supra n 2, at pp 135-6.

Reflection point: Law practices should explore how to reimagine their business processes given that AI technologies may increasingly take over routine legal tasks (see our November 2018 article titled “Lawyer + Machine = ?”).

AI as Legal Revolution

Although much attention has been focused on the impact of AI technologies on the practice of law, a much bigger question on the horizon for lawyers is the effect of AI on national laws. In Robot Rules: Regulating Artificial Intelligence, Jacob Turner of Fountain Court Chambers suggests that AI will pose new challenges for existing legal systems not only in the ethical realm, but also in determining legal responsibility (e.g. who is responsible if AI were to cause harm?) and rights (e.g. should AI be granted a separate legal personality?).13 Jacob Turner, Robot Rules: Regulating Artificial Intelligence (Cham, Switzerland: Palgrave Macmillan, 2019), at p 37 (“Robot Rules”).

Some commentators have argued that new laws are unnecessary to address the AI challenge, as an incremental development of existing laws would suffice.14Robot Rules, id, at pp 41-2. However, Jacob Turner observes that the uniqueness of AI, such as AI’s ability to reinvent itself through machine learning, makes AI “legally independent” of humans and requires a rethink of the fundamental legal concepts of agency and causation.15Robot Rules, supra n 13, at pp 57, 64 and 70-78.

Even if the conceptual legal issues surrounding AI are murky, a recent joint law reform consultation paper (spanning more than 200 pages) issued by the UK and Scottish Law Commissions on automated vehicles16 Law Commission and Scottish Law Commission, Automated Vehicles: A Joint Preliminary Consultation Paper (Law Commission Consultation Paper No 240; Scottish Law Commission Discussion Paper No 166) (8 November 2018) <https://www.lawcom.gov.uk/project/automated-vehicles/> (accessed 28 November 2018). suggests that AI’s impact on current national laws will be significant. For example, in the chapter on criminal liability, the Commissions examined in great detail novel questions such as whether current offences were incompatible with automated driving, whether a new system of sanctions for automated vehicles should be imposed and whether new offences relating to automated vehicles should be created.17Id, at pp 125-152.

Reflection point: Lawyers should consider playing an active role in potential law reforms pertaining to AI, as AI is too complex a subject to be limited to academics and policy-makers.

Conclusion

Professor Hawking’s real concern was not that AI will outsmart humans, but that when that day comes, we must ensure that AI’s goals will be aligned with ours, as otherwise humankind would be at risk of being displaced by a super-intelligent AI.18Brief Answers, supra n 1, at pp 184, 188. While lawyers do not need to deal with such big questions, they face the same challenge of whether they should allow AI to shape their legal practice, or whether they should shape AI to align with their goals of legal practice.

Endnotes   [ + ]

1. Stephen Hawking, Brief Answers to the Big Questions (London: John Murray (Publishers), 2018), at p 184 (“Brief Answers”).
2. See e.g. Thomas H. Davenport, The AI Advantage: How to Put the Artificial Intelligence Revolution to Work (Cambridge, Massachusetts: The MIT Press, 2018), at pp 6-7 (“The AI Advantage”).
3.The AI Advantage, id, at pp 6-8.
4.The AI Advantage, supra n 2, at p 7.
5.The AI Advantage, supra n 2, at pp 11, 14 and 16.
6. Nigel M. de S. Cameron, Will Robots Take Your Job?: A Plea for Consensus (Maiden, MA: Polity Press, 2017), at p 30.
7.Id.
8.The AI Advantage, supra n 2, at p 130.
9. John Paul Mueller and Luca Massaron, Artificial Intelligence for Dummies (Hoboken, NJ: John Wiley & Sons, Inc., 2018), at p 218.
10.The AI Advantage, supra n 2, at p 132.
11.The AI Advantage, supra n 2, at p 133.
12.The AI Advantage, supra n 2, at pp 135-6.
13. Jacob Turner, Robot Rules: Regulating Artificial Intelligence (Cham, Switzerland: Palgrave Macmillan, 2019), at p 37 (“Robot Rules”).
14.Robot Rules, id, at pp 41-2.
15.Robot Rules, supra n 13, at pp 57, 64 and 70-78.
16. Law Commission and Scottish Law Commission, Automated Vehicles: A Joint Preliminary Consultation Paper (Law Commission Consultation Paper No 240; Scottish Law Commission Discussion Paper No 166) (8 November 2018) <https://www.lawcom.gov.uk/project/automated-vehicles/> (accessed 28 November 2018).
17.Id, at pp 125-152.
18.Brief Answers, supra n 1, at pp 184, 188.

Director, Legal Research & Development
The Law Society of Singapore
E-mail: alvinchen@lawsoc.org.sg

Executive Officer, Legal Research & Development
The Law Society of Singapore
E-mail: stellachen@lawsoc.org.sg