First, they started talking, then they started learning, now they are automated—are machines taking over? Is man creating “man” out of machines by the breath of algorithms? The ever-advancing and disruptive innovations of AI have permeated virtually all fields of life, including the legal profession that has largely remained conservative since the 1800s.1 This essay appraises the potential and present realities of AI under the Nigerian legal profession, its sustainability, and the challenges therein.
Inter alia, it finds that AI is at a sprouting stage in Nigeria, a fortiori, the legal profession,largely due to poor datasets and existing socio-legal constraints. And that even in its most realistic automated design, it may not be capable (at least not in the nearest future) of eliminating completely the complex and creative judgments of “a human [lawyer] in the loop.”2
Finally, it concludes that any sustainable role of AI in law practice in Nigeria is/will be predominantly complementary and/or indirect through the legal issues it may instigate and/or an alternative court system or dispute resolution mechanism.
No one definition of AI has a universal acceptation.3 Coined in 1956 by John McCarthy as “the science and engineering of making intelligent machines,” AI subsumes all technologies that enable computers to undertake intelligent functions, otherwise capable of only being performed by humans.4 Thus, machine learning, deep learning, robotics, internet of things/bodiesare all among aspects of AI. While its classification varies, it may be classified as “Weak AI” – i.e. designed to perform specific tasks, and “Strong AI” – general-purpose AI. The former is already with us, while the latter is still an illusion considered a dystopia by many.
The operation of AI tools centres on the recognition of patterns of cognitive behaviour from big datasets encrypted into them using machine language, that is, 0 and 1. Thus, the larger the datasets, the better the functionality of the devices. Hence, access to large and accurate datasets underpin its workability.5 AI is gradually being used in adjudication. In the American case of Wisconsin v Loomis,6 it was relied on to sentence the defendant.
The Case for AI under Nigerian Law: Futuristic Projections
Despite its isolationist inclination, the legal practice in Nigeria is opening up to the use of AI. Although statistics are hardly available, it may be submitted that most, if not all, lawyers in Nigeria use “weak AI”— e-mail, Google search, and other social media tools in practice.
This submission is inspired by a recent survey by the American Bar Association (ABA)7which found that 90% of American lawyers use online research tools, and 73% of firms and 79% of lawyers are on social media for professional purposes (with LinkedIn and Facebook leading with 82% and 47% respectively). Also, while 10% already use “advance” AI tools, 36% think it will become mainstream in the near future.Similarly, according to a Deloitte study,100,000 legal roles will be automated by 2036.8
The tendril of legal AI in Nigeria gained unequalled spirals with the unveiling of LawPavilionPrime – a tool with law reporting/predictive capabilities and TIMI – the first true AI chatbot that assists lawyers with procedural laws.9 The ongoing Covid-19 pandemic has further strengthened the case for AI in Nigeria. Although there are now remote hearing of cases in some States, e.g. Lagos, it is submitted that this would have been more effective had AI been entrenched in Nigerian law practice.
Situations such as the above, perhaps justify the “outcome thinking”10of AI adventurists – a future of online Courts with automated judges and lawyers. Even the judiciary may be open to this eventuality. In Lola v Skadden11it was held to the effect that an activity capable of being performed entirely by a machine is not akin to being engaged in the practice of law, though the activity is one that should ordinarily be performed by a lawyer properly so called.
The Case Against AI under Nigerian Law: Actual Realities
It has been argued that any thought of AI in law practice in Nigeria is a mere legal el dorado. For it may not be rightly said that AI exists in Nigeria, or elsewhere today. Thus, “AI is whatever hasn’t been done yet.”12This notion of AI, which is inclined to “Strong AI” and seems to ignore or gloss over “Weak AI” may suggest that most of the available literature on AI tend to concentrate on the potentials rather than the realities.
So that a tool like Timi which is acclaimed for intelligent acts may not be intelligent after all. For according to Surden, “…today’s AI [Including Timi]is not a thinking machine” but a heuristic machine that uses a computational approximation to detect patterns in data. And the term “learning” associated with AI is at best functional.13
In terms of this essay, perhaps this argument raises a considerable conundrum. How can that which does not exist be capable of creating a sustainable law practice in Nigeria? Whatever the case, this writer is inclined to think that this argument is quite simplistic.
The Role of AI in Sustaining Law Practice in Nigeria
1. Complementary/Symbiotic Role
AI can assist lawyers/courts/bailiffs to be more effective and efficient in the following ways:
(a) E-Document Recovery/Contract Review:
AI tools can review large volumes of documents/contracts in seconds. Thus maximising time, especially in company restructuring and large-documents-prone litigations, e.g. election petitions.
(b) Legal Research: AI tools like Timi can assist with research on procedural laws. Also, issuance/filing/service of court processes can be more efficient with e-mail.
(c) Risk/Case Management/Analysis:
AI can reduce the rate of frivolous suits that inundate our courts. Tools like Ravellaw can analyse and predict the outcomes of cases.
(d) Due diligence:
AI can help in regulatory compliance, e.g. Kira can assist in due diligence.Also, AI can ensure mass access to courts/justice through pro bono bot lawyers like DoNotPay.
(e) Judicial Decisions:
AI can assist judges to dispense justice expeditiously and reduce the backlog of cases in courts. In the US tools like COMPAS assist judges in sentencing.
(f) Much time is lost to delivering/reading judgments in courts. Legal chatbots could be used instead, especially in civil matters.
2. Continual Litigation Inflow Role
Unlike in law, AI is thriving in other areas in Nigeria, e.g. finance/banking.14 Hence, whether AI is particularly used in law practice or in other fields, it has the promise of sustaining law practice through the legal issues it is likely to generate in such areas as privacy, intellectual property (IP), tortious liability etcetera.
3. Alternative Court System/Dispute Resolution Mechanism
“Strong legalbots” (advance AI) could become an alternative to conventional court system in themselves, thus multiplying the merits of ADR.
Evidential/Legal and Other Challenges
(i) Legal Constraints:
AI devices, being intellectual properties, are protected by trade secrets which cannot be legally called for in evidence, as was seen in the US case of Washington v Fair.15Thus, courts accept outcomes they don’t know how they are reached.16 The solution to this may be to adopt one of the proposals of France’s national AI strategy which aims to make AI algorithms open, thereby eliminating the trade secret question.17
(ii) Liability Distribution:
Drawing from the malfunction of the Microsoft chatbot, Tay, which exhibited the character and capacity to use slur, racial and anti-Semitic languages,18 it is not inconceivable that an AI lawyer may be in contempt of court. Who will be liable for such contempt, the lawyer or its creator?
(iii) Mechanical Infallibility/Neutrality:
In assisting Courts to arrive at decisions, AI may present results that are extraordinarily objective to be “true.” In this case, the position in Argentina, where courts may adopt, rewrite or reject the results of AI tools may be adopted.19
(iv) Human-to-machine Prejudice:
The natural biases of the creator may be transferred to an AI tool. For instance, ProPublica reported that the COMPAS (an AI tool with the capacity to predict the possibility of a defendant reoffending) is highly biased against black defendants as being likely to re-offend than their white counterparts.20 Considering the mutual tribal suspicions among the tribes in Nigeria which often infiltrate national database (e.g. manipulation of census figures), it may not be impossible to have legal AI tools that unwittingly reflect this character.
(v) Lack of Government Interest/Inadequate Access to Accurate Datasets:
Inadequate access to datasets plus government’s interest in advancing AI technologies related to security, healthcare and agriculture, but not necessarily law until after 2030,21 may affect AI’s role in sustaining legal practice in Nigeria.
(vi) Save certain routine tasks, law is largely cognitive-based.
Sustainability is achieved where merits are maximised and demerits minimised. Herein lies the sustainable future of law practice in Nigeria. The lawyer-machine relationship in furtherance of justice. AI is not a usurper; it does not, in essence, seek to take lawyers out but to bring machines in for more efficient, accessible and expeditious administration of justice. It is akin to an impossibility becoming possible—one throne, two kings; one profession, two practitioners.
*Bizibrains Okpeh is a writer, a disability rights advocate, and a legal practitioner in Nigeria. You can reach him on firstname.lastname@example.org or 07061096037
1. Bernard Mar, The Future of lawyers: Legal, Tech, AI, Big Data And Online Courts, available atcom/sites/bernardmarr/2020/01/17/the-future-of-lawyers-legal-tech-ai-big-data-and-online-courts/ [accessed 2/6/2020]
2. Harry Surden, Artificial Intelligenc, and Law: An Overview, p.1321, available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3411869 [accessed 2/6/2020]
3. Cheldreth et al, AI in Nigeria p.2, available at https://www.google.com/url?q=https://research.google/pubs/pub48985.pdf&sa=U&ved=2ahUKEwi40rCfuO3pAhWH3KQKHdB5DSwQFjAAegQIAxAB&usg=AOvVaw0WCtgKwD0Wbh4h2kvzgx2x [accessed 4/6/2020]
4. Mirjana Stankovic et al, Exploring Legal, Ethical and Policy Implications of Artificial Intelligence p.5, available at https://www.researchgate.net/publication/320826467_Exploring_Legal_Ethical_and_Policy_Implications_of_Artificial_Intelligence [accessed 2/6/2020]
5. Cheldreth, (n3) p.10
6. In Stankovic, (n4) p.28
7. ABA Profile of the Legal Profession, 2019 pp.52-54, available at https://www.americanbar.org/content/dam/aba/images/news/2019/08/ProfileOfProfession-total-hi.pdf [accessed 4/6/2020]
8. Charlotte Truman, How AI is Impacting the UK’s Legal Sector, available at https://www.computerworld.com/article/3412357/how-ai-is-impacting-the-uks-legal-sector.html [accessed 4/6/2020]
9. Ademola Adeyoju, Artificial Intelligence and the Future of Law Practice in Africa, available athttps://www.tekedia.com/artificial-intelligence-and-the-future-of-law-practice-in-africa1/ [accessed 3/6/2020]
10. Mar, (n1)
11. In Adeyoju, (n9)
12. Tesler’s Theorem, quoted in Mark Aloof, Artificial Intelligence: An Introduction, available at http://people.cs.georgetown.edu/~maloof/cosc270.f17/cosc270-intro-handout.pdf See also https://en.m.wikipedia.org/wiki/Artificial_intelligence [accessed 7/6/2020]
13. Surden, (n2) pp.1310-11
14. Cheldreth, (n3) p.3
15. Lael Henterly, The Troubling Trial of Emanuel Fair
https://www.seattleweekly.com/news/the-troubling-trial-of-emanuel-fair/ [accessed 6/6/2020]
16. Stankovic, (n4) p.28
17. Ivanov et al (eds), International and Social Impact of Artificial Intelligence, p.47 https://russiancouncil.ru/papers/AI-Paper44-Eng.pdf [accessed 6/6/2020]
18. Stankovic, (n4) pp.29-30
19. Franco Diangana, Victor Frankenstein’s Responsibility? Determining AI Legal Liability in Latin America, p.168 https://giswatch.org/sites/default/files/gisw2019_artificial_intelligence.pdf [accessed 6/6/2020]
20. Stankovic, (n4) p.28
21. Cheldreth, (n3) pp.3-4