INTRODUCTION

On the 3^rd of July, 2020, Oladapo Daniel Oyebanjo popularly known by his stage name, “D”Banj” filed a case against Seyitan Babatayo for a claim of 1.5 Billion Naira through his counsel Mike Ozekhome (SAN) at the High Court of the Federal Capital Territory, Abuja. ¹ What however got the attention was not the 1.5 Billion Naira Claim but the manner in which the claim was written in the court process. The 1.5 billion Naira claim was divided into two claims, a 1 billion Naira, general damages and a 500 million Naira, aggravated, punitive and exemplary damages. However, the former claim was written as “One Billion Million Naira”. ²

The error in the claim became a trending topic on “Twitter Nigeria” on the same day that is, 3^rd of July 2020. At around 10 o”clock in the evening, “Mike Ozekhome” and “One Billion Million” were trending topics 6 and 14 on the trends of Twitter Nigeria. ³ The fact that the process emanated from the office of a well-known Senior Advocate of Nigeria shows the infallibility of man no matter how highly experienced or dutiful s/he is. This is where Information and Communication Technology (ICT) comes in.

The processing power of computer systems, the ubiquity of the internet and the cognitive abilities of Artificial Intelligence have made ICT applicable in all spheres of life. However, the legal sector has been slow in incorporating high end ICT, in fact Martin remarks that transactional lawyers have been “notoriously slow” in integrating technology in their practice. ⁴ This reality is worse in Nigeria where the biggest arguable use of technology in legal practice is the automated services of some states” land registry and the yet to be efficiently running platform of the Corporate Affairs Commission. ⁵

In advanced countries, the trend has changed in recent years as there has been a massive use of ICT in legal practice. Of course, this discourse is not about using word processing programmes to make legal documents but rather application of advanced ICT tools famously termed “Artificial Intelligence (AI)”. The concept of AI has been defined as using computers to automate task normally require human intelligence. ⁶ It has also been defined as emphasising technology that automates specific type of task thought to require intelligence when people perform them. ⁷ Thus, AI transcends using computer system to perform tasks but goes a step further for the computer process data using cognitive methods to perform tasks that will ordinarily involve human intelligence.

Lately, there is a new branch of research termed “AI and Law” which focuses on the application of AI to law. More comprehensively, Surden defines AI and Law as “the application of computer and mathematical techniques to make law more understandable, manageable, useful, accessible, or predictable”. ⁸ Simply put, AI and Law deals with computer methods used in the practice of law or in the process of information used in legal practice. There are many of such methods, in fact, 35 of such applicable methods in law is identified. ⁹

One of these methods is “legal document automation” which is the process of using AI to draft legal documents. ¹⁰ This paper is a theoretical analysis of a simple programme written by the first author to draft simple land sale agreement which will fulfil the basic requirements of a land sale agreement in Nigeria. As much as this paper may sound technically as a scientific exploit than a legal research, this is a research in the emerging field of AI and Law and fittingly a legal research. The paper will exploit the tenets of legal document automation, exploring the technical, ethical and legal issues that may arise in the development and use of this AI method. The paper will also examine the essentials of a land sale agreement before justifying the compliance of the programme to these essentials. It will also examine how the programme addresses the already discussed technical, ethical and legal issues and will conclude on the prospects of the programme and other similar programmes in Nigerian legal practice. Technical details of the programme will be explained also to elucidate on the functionality of the programme.

LEGAL DOCUMENT AUTOMATION

As explained above, legal document automation is one of the methods of AI and Law which involve the use of AI to make competent legal documents. The method is one of the earliest methods of AI and Law, Betts and Jaep in an article highlight the history of legal document automation particularly contract drafting softwares. ¹¹ Word processing software in the 1980s and 1990s was arguably the emergence of legal document automation as this software offered simple computer file management systems, automatic numbering tools, basic “master documents,” macros, and document “merge” functions. ¹² With the use of word processing software, drafting of legal document became less tedious, cheaper, more organised and productive as less time was required in making similar documents. ¹³

As earlier emphasised in the introduction section, AI and Law transcends merely processing information with the use of computers, the processing must involve intelligence for it to qualify as an AI method. Thus, it can be argued that the use of word processing software in the 1980s and 1990s did not qualify as a method of AI and Law. However, the development of “expert drafting systems” in the latter years of 1990s whereby a document is produced using “logic tree” from a series of answer supplied to the system ¹⁴ satisfies the requirement of intelligence and the beginning of legal document automation.

These expert drafting systems however were critiqued for their text being hard coded and inflexible and preventing the user from making changes after the production of the document. ¹⁵ “Document assembly engines” quickly followed the drafting systems and these engines allowed for modification and combination of drafting templates with other texts. ¹⁶ As much as this allowed for more flexibility, it takes away the speed and ease offered by expert drafting systems.

Machine learning has provided for the betterment of these programmes ¹⁷ ; however, machine learning is not a focus of this paper. The use of these programmes has generated some technical, ethical and legal issues which the paper will thereafter explore.

TECHNICAL, ETHICAL AND LEGAL ISSUES GENERATED BY LEGAL DOCUMENT AUTOMATION

Although there are three different categorisations of the issues raised by the use of legal document automation, that is technical, ethical and legal issues. All these issues will be addressed under one heading as these issues just like the technology itself are newly emerging. Some of these issues may also not be strictly tailored under any of these categorisations and may reflect in two or even all of the categorisations.

ESSENTIALS OF A LAND SALE AGREEMENT IN NIGERIA

The main subject of this paper is a programme written by the first author that automatically generates a land sale agreement using a few essential facts which can be inputted by either a lawyer or a client that needs the document. Before an analysis is made into the technicalities and functionality of the programme, it is pertinent to examine the rudiments an agreement must fulfil to qualify as a basic land sale agreement.

Like most simple contracts in Nigeria, a land sale agreement is governed by the common law. Elementarily, this means that the contract must have offer, acceptance, consideration, intention to enter legal relationship and capacity to contract. ³³ These are facts that may not be apparent on the face of the agreement though there may be pointers in the face of the agreement from which some of these facts may be deduced. There are, however some more specific requirements that deals with the form of the agreement.

• Agreement must be in writing and signed:

By virtue of the Statutes of Fraud, ³⁴ all contract for sale or any other disposition of land or interest in land must be in a written memorandum or note and must be signed by the party to be charged or through an authorised person. ³⁵ The Statute of Fraud has been superseded in many states of the federation by specific legislations regulating though it should be noted the requirement of the contract being contained in a note or memorandum and same must be signed by the party or the authorised person instead. ³⁶

It should be, however noted that these legislations are silent about the form in which the agreement should take. The Court of Appeal in the case of Adeniran v Olagunju ³⁷ stated per Justice Amaizu J.C.A that;

“I have mentioned earlier that the law requires the evidence of a transaction in a sale of land to be in a note or a memorandum. It is necessary to mention also that no special form is prescribed for the note or memorandum. However, from the authority of decided cases particularly in Hamilton v Kofi Mensah (1937) 3 WACA 224 a document that will satisfy the description of a note or memorandum within the meaning of the statutory provision must contain the following details-

• names of the parties or enough description of the parties.

• the ample description of the subject matter of the contract.

• the consideration for the alleged contract and

• the document must contain the signature of the party to be charged or that of his agent or a representative duly authorised by him.” ³⁸

The interpretation given to the requirement of an agreement for land sale agreement under the Statute of Fraud and other similar legislation is quite succinct. Although, it should be emphasised that most lawyer adopt a style that provide the names and address of the parties, a clear description of the property, the consideration, signature of all the parties to the agreement and at least two attesters to the agreement.

• Illiterate jurat where one of the parties is an illiterate:

The law usually protect illiterate who are parties to an agreement to ensure the illiterate understand the implications of the agreement they are entering into thereby protecting the illiterate from possible fraud. ³⁹ This protection was made compulsory for any document executed by an illiterate. ⁴⁰ The form this jurat is to take is to state the name and address of the person who makes the interpretation, a clause to show that the illiterate understands the content of the document before the execution of the document. ⁴¹

• Agreement must be prepared by a legal practitioner

A land sale agreement or any document for sale or transfer of land is required to be prepared by a legal practitioner by virtue of Land Instrument Preparation Laws of states in Nigeria ⁴² as well as the Legal Practitioners Act. ⁴³ This poses a problem for document for sale or transfer of property made with AI. A document made by AI is definitely not made by a legal practitioner and may fall short of this requirement. There is however the argument that if the legal document automation was used or authorised by a legal practitioner, then the lawyer can be said to have franked the document especially if his signature is appended on the AI generated document.

• Other conventional miscellaneous terms:

Although these terms do not directly affect the validity or form of a land sale agreement, they have become conventional in drafting to include them in any land sale agreement, they include;

• Recitals which are not operational part of the agreement but an explanatory note to explain the intent of the parties in entering the agreement.

• An indemnity clause to indemnify the purchaser(s) against any third-party claim on the property.

• An optional deposit clause where the parties agree for the payment of a deposit that is liable to be forfeited if the purchaser defaults. ⁴⁴

• A clause stipulating that the vendor(s) undertake to execute any document to perfect the assignment of right in the property to the purchaser(s).

REVIEW OF THE BASIC LAND SALE AGREEMENT SOFTWARE

The programme was written fully with Python ⁴⁵ language specifically version 3.81. some internal and external modules were used in the implementation of the programme. The datetime internal module, specifically the date function of the module was used in specifying the execution date of an agreement. The choice function of the random module was also used to give the programme a bit of flexibility.

Python does not have a module or function to directly manipulate word document files ⁴⁶ , therefore an external library was installed using “pip installer” ⁴⁷ . The external library installed is the python-docx ⁴⁸ which is a library for creating and updating Microsoft Word (.docx) files.

The Document function of the docx module of the python-docx library was used extensively in the programme to create word document files while the WD_ALIGN_PARAGRAPH and WD_LINE_SPACING functions from the docx.enum.text ; the WD_STYLE_TYPE of the docx.enum.style and the Pt function of the docx.shared module were used in styling and structuring the text generated by the programme.

The programme used basic logic, loops and conditionals to instruct the programme in manipulating results from inputs in creating a programme that can qualify as intelligent. The first author was also careful in naming variables used in the programme so it will be easily understandable for any future Pythonista or Pythoneer that may work to improve the programmer in the future. It should be pointed out that the programme can only produce documents between natural persons.

The programme being a code that will require several inputs before execution was written without the utilisation of defined classes but the programme utilised a lot of function to effectively remedy some inadequacies in the custom internal modules, methods and functions of Python.

CONCLUSION AND RECOMMENDATIONS

It is imminent that most part of legal practice will become obsolete in the nearest future as intelligent computer programmes will be able to render these services using less time and less expensive resources. Computing or programming is not a conventionally part of the legal practice, however like natural languages, it takes only understanding, discipline and practice to be proficient in it.

The reviewed programme shows the prospect of AI in law and how the legal practice can be bettered with the acceptance of these technologies. If the incorporation of AI is imminent, lawyers and legal researchers should not be mere spectators or users of these technologies. As explained above, the right in these programmes are governed by the law of contract and intellectual property and it is likely to be vested in computer programmers than lawyers. Thus, in the future it is very possible that the right to most legal services will be owned by computer programmers if the tide is not changed and lawyers and legal researchers take over the responsibilities of writing programmes for use in the legal practice.

It should emphasise that AI and Law is an established field of research in many universities in some countries, Nigerian universities are however still stuck in the teaching and research of substantive law, there should be a paradigm shift and some of the researches should be focus on the practical application of ICT especially AI in law. Flowing from this narrative, this paper hereby makes the following recommendations:

• Legal researchers should be encouraged to focus on technical and practical application of ICT to legal practice.

• The field of AI and Law should be incorporated into the curriculum of law more appropriately at post-graduate level with law graduate student with interest in programming, students from law related courses with intermediate or advanced knowledge in programming language(s) and graduate student of computer science, software engineering or similar courses with interest in making AI tools for legal practice should be target students. Also, Legal scholars and programming experts should draw a comprehensive curriculum for the field of AI and Law for adoption of by the Nigerian University Commission.

• Faculties of law should encourage undergraduate to choose a career in the field of AI and Law or in the application of ICT in legal practice by incorporating it in the extra-curriculum activities of the undergraduate by floating ICT Clubs, competitions, periodic lectures and symposiums and student conferences on ICT related aspect of law.

• Lawyers should accept and encourage the use of ICT tools in their legal practice and should be mindful of the implications of agreements they make for development of ICT tools on legal practice.

• The Body of Benchers, the Bar Council as well as the Nigerian Bar Association should solicit for the enactment of legislation that will regulate and restrict the interest of non-lawyers in AI tools that provide legal services.