Written by Cindy Chua (Associate Author) | Mentored by Josh Lee | Reviewed by Joey Pang
LawTech.Asia is proud to conclude the third run of its popular Associate Author (2020) Programme. The aim of the Associate Authorship Programme is to develop the knowledge and exposure of student writers in the domains of law and technology, while providing them with mentorship from LawTech.Asia’s writers and tailored guidance from a respected industry mentor.
In partnership with the National University of Singapore’s alt+law and Singapore Management University’s Legal Innovation and Technology Club, five students were selected as Associate Authors. This piece, written by Cindy Chua and reviewed by industry reviewer Joey Pang (DBS), marks the fifth and final thought piece in this series. It examines the various issues arising from the use of commonly seen legal technologies in Singapore, and conducts an international scan of legal technology regulatory pictures in several jurisdictions before proposing a potential way forward for Singapore.
With the rise of legal technology (“legal tech”), the boundaries of regulation are constantly being pushed by new frontiers of innovation. Disruptive innovation is typically fast-paced and exponential, so it remains a challenge for regulations to sufficiently address legal technology. Concurrently, it is noted that a lack of transparency and clarity in regulation of legal technology can have the undesirable effect of deterring innovation. As such, to promote the development of legal tech, it is important to have a clear regulatory framework that “ensure[s] the safety of users and the public”, while “facilitating the commercial use and consumer enjoyment of innovation”.
This paper discusses regulating client-facing legal technology. Client in “client-facing” refers to the client in a client-solicitor relationship. There is a need to protect clients as the technology is used to serve their legal needs. In particular, the focus is placed on some of the most prevalent types of client-facing legal tech, namely legal chatbots, self-service automated document generation, and marketplaces. In order to implement effective regulations for the client-facing legal tech, the core functions of the different types of client-facing legal technologies are first examined. After understanding the purpose behind each type of legal technology, existing legal or regulatory mechanisms that may already apply in relation to the use of such legal technology systems, and how they may apply in each case, are then examined. The caveat here is that these analyses should not be taken to mean that legal technologies should definitely be regulated through these levers, but rather, this paper offers these as possible lenses that users of these legal technologies should think about when using deploying legal tech. In the final part of this article, the question of whether there should be a one-size-fits-all regulation introduced for legal technology will also be considered.
Regulation of client-facing technologies
Key distinction: Legal information vs legal advice
We begin with chatbots and self-service automated document generation. The key value of these legal technologies is the provision of legal information. Legal information is to be distinguished from legal advice. Although “legal information” is not defined in Singapore law, the difference between legal information and legal advice is that legal information involves stating a general point of law (i.e. what is the law), versus legal advice which is concerned about the outcome of applying the law against the facts of a particular situation. In practice, the distinction between the two can be murky, and much depends on the circumstances of each case.
Chatbots allow users to enter their queries and receive an immediate answer from the chatbot (often without further human assistance). Responses are based on “decision-trees” that the law firms have programmed the chatbots with. This section will limit the discussion on the chatbots to chatbots that use discrete decision-trees (i.e. those that provide a limited number of options as responses), and not conversational AI virtual assistants similar to Siri or Alexa that use stochastic decision-making models (and may also learn from real-life user input). This distinction is drawn because the majority of chatbots in the legal industry still utilise decision-trees. This section also makes the assumption that the chatbot is made by a law firm.
Legal chatbots providing legal information, with the appropriate disclaimers, are unlikely to be giving legal advice. At the outset, it is posited that identifying the line between legal advice and legal information depends on the specificity of a response (from a lawyer or legally-trained person, or the “provider”), as guided by (i) the input provided from the person seeking the legal advice/information (the “seeker”); and (ii) the nature of any representation(s) made by the provider to the seeker. With decision-tree chatbots today mostly suggesting that seekers could explore certain options or broadly expect certain outcomes, along with the fact that seekers are generally only able to input a limited range of information, as well as appropriate disclaimers made by the chatbot system owners, it is hard to imagine chatbots being considered as giving legal advice (despite the absence of judicial decisions to this effect).
On this basis, would there be liability for legal information that is provided incorrectly? Legal advice given by lawyers is presently regulated by s 5 of the Legal Profession (Professional Conduct) Rules 2015 (“PCR”). There is, however, no corresponding provision for legal information, even if the information is wrong or inaccurate. In any event, it is noted that even if the chatbot gives legal advice, the response from the chatbot may fall under a lacuna in the law as there is presently no statutory provisions nor judicial precedents governing legal advice from chatbots. In short, there are no legal industry-specific regulatory levers yet for chatbots that provide inaccurate legal information.
Nevertheless, an aggrieved user who had acted on the chatbot’s output could pursue a claim on other legal bases, such as misrepresentation. For fraudulent misrepresentation, it will require fraudulent intent on the part of the chatbot operator. Since the chatbot operator would refer to the company, the state of mind would be that of the person who configured the algorithm from the point of programming up to the point that the relevant contract was formed, as held in Quoine Pte Ltd v B2C2 Ltd. It is posited that such a claim would, in most cases, be unsuccessful, given the high legal threshold requiring strong evidence of dishonest intent.
Another cause of action available to an aggrieved user could be under the tort of negligence, whether negligence is made out would depend on whether the chatbot operator owed a duty of care to the individual using the chatbot. This would in turn be determined by whether the loss was reasonably foreseeable, whether there was sufficient proximity between the chatbot operator and the individual using the chatbot, and whether there was a voluntary assumption of responsibility by the chatbot operator.
In a hypothetical situation where the chatbot user relies on the output given by the chatbot and suffers a loss, factual foreseeability would only be fulfilled if it was foreseeable that the advice given by the chatbot could have resulted in the harm suffered by the chatbot user. Proximity is likely to be satisfied – there is causal proximity since the chatbot operator’s acts would likely have influenced the chatbot user. As for voluntary assumption of responsibility, this will depend largely on whether there are sufficient disclaimers in place and whether such disclaimers put a user sufficiently on notice, as otherwise it would not be difficult to relate to chatbot users forming an expectation that they can rely on the responses given by the chatbot. Within the specific context of chatbots offered by law firms, i.e. law firms as chatbot operators, from a policy perspective, it may be that more will be expected from law firms as there is an expectation that lawyers and law firms should strive to provide accurate legal advice. Thus, law firms and legal practitioners would be well advised to ensure that any chatbot operated are accurate or sufficiently disclaimed.
Automated document generation
Automated document generation tools generally allow legal documents to be produced after a user provides relevant information (often by answering a set of guiding questions). Such tools are often used in two ways. First, a lawyer could use the tool to draft documents for her client. Second, a legal services consumer could use the tool as a form of “self-service”, with the option of having a lawyer vet the contract thereafter (usually with the payment of an additional fee). Examples of automated document generation tools include Zegal and Contract Express.
Under the first scenario (where a lawyer uses the tool to draft software for her client), assuming that a practitioner-client relationship exists, r 5 of the PCR would apply. Hence, the lawyer must be diligent in her advice given to the client, even if document automation software is used. Thus, it is likely that a lawyer using the software would be held to the same standard of diligence in her advice given to the client.
Under the second scenario where no lawyers are involved, it remains unclear whether offering tools to generate their own documents constitutes a provision of legal services. Some observations may be made. First, arguments could be made both ways about whether the self-service use of document generation tools constitutes the provision of legal services. On one hand, it could be argued that just like the use of templates on Word or the sourcing of contract templates online, the mere use of an automated contract generation tool would not be considered as legal services, especially when there is (as is often the case) no contract of retainer signed. On the other hand, it could be argued that the tool constitutes the agglomeration of legal expertise of the lawyers in the law firm within an easy-to-use platform, and when applied to a self-service user’s specific needs, could be considered the provision of a legal service. Further, there may also be reliance placed on the fact that the tool was provided within the context of a practitioner-client relationship. Nonetheless, the jury remains out there on this issue. Second, as mentioned above, the LPA and PCR regulate the provision of legal services by qualified natural persons (and not legal persons). This means that where an automated contract generation tool was offered by a law firm, even if the use of the tool were considered the provision of legal services, such a law firm may not be liable under the LPA and/or PCR, although whether this corporate veil argument is robust remains questionable.
In short, unless a lawyer was specifically involved in the drafting or vetting of the contract, it is remains an open question whether the use of such automated document generation tools is regulated by the LPA or the PCR. Accordingly, it is posited that regulatory intervention would be helpful to clarify the situation and better protect self-use consumers of automated document generation tools.
Online marketplaces provide a platform for lawyers and law firms to advertise their legal services online. This allows people who are looking for legal advice and/or representation to find lawyers using just one platform. Prominent examples of such online marketplaces in Singapore are Asia Law Network and SingaporeLegalAdvice.com.
The essential issue about online marketplaces is advertising legal services. Marketplaces are not subject to the same standards as lawyers. In contrast, Part 5 of the PCR regulates touting and publicity by lawyers, with the aim of upholding the standards and honour of the legal profession. Although the online marketplaces themselves may not be subject to the rules under the LPA or the PCR, the question is whether lawyers and law firms using the online marketplace services would fall afoul of the PCR.
A lawyer’s general responsibilities relating to publicity are found under r 42 of the PCR. Under r 2(1) of the PCR, publicity includes advertisements communicated through the Internet. Thus, online marketplaces would be captured under r 42, so lawyers and law firms using the online marketplace services would still be subject to the same publicity requirements under r 42 of the PCR.
Rule 44(1)(a) prohibits lawyers from conducting publicity in a manner which is likely to “diminish public confidence in the legal profession or … bring the legal profession into disrepute”. In particular, whether the publicity would diminish public confidence depends on the “public expectation of lawyers”. As Prof Jeffrey Pinsler SC points out, the public would have the expectation that lawyers will “act honourably and honestly and with integrity and dignity”.
Thus, the use of marketplaces per se would not be an issue under the PCR insofar as the relevant rules under Part 5 of the Legal Profession (Professional Conduct) Rules 2015 are complied with.
Data protection concerns
Aside from the regulatory areas already raised by the LPA and PCR, other areas may also require greater regulatory scrutiny in respect of legal practice. One such area is personal data protection. Legal technology tools, be they tools that utilise more advanced technologies (such as predictive analytics, natural language processing, or machine learning) or tools that utilise baseline technologies (e.g. basic functionalities of practice management software), often process or handle personal data. This raises some data protection concerns, but the data is offered some degree of protection by the lawyer’s duty of confidentiality. On top of this duty, contractual agreements and the PDPA may provide additional protection.
As a preliminary point, regulation sometimes serves as a consumer protection safeguard. Since how the LPA affects legal technology is a novel area of law, there remains many uncertainties in this area of law. Providing clarity in regulations is important as “a lack of transparency and control might prevent [legal service consumers] from fully exercising their rights”.
With the wide range of legal technology tools in the market and consistent innovations, it would be interesting to consider whether there is a possibility of a one-size-fits-all regulation that regulates all legal technology use cases. Given the diversity of legal technology tools, it would be difficult to directly regulate the technology enabling each tool. Rather, it would be more realistic to regulate the outcomes of these tools. Some policy objectives that could serve as guiding points in regulating outcomes include: (1) preserving the ability to provide legal advice to lawyers to ensure that only those appropriately skilled and trained can advise people on their legal rights and obligations; (2) having publicity rules to preserve the nobility and high standing of the profession; and (3) to regulate privilege and personal data issues to preserve the high level of trust inherent in the client-lawyer relationship.
Since many of the outcomes that we seek to achieve with the use of legal technology are in line with the principles underpinning the regulation of the legal profession, it is argued that a possible approach to regulating legal technology tools would be to continue relying on the LPA or PCR (rather than creating a new regulation altogether), with some potential adjustments to address legal technology-specific issues. Such an outcomes-based, technology-neutral approach in regulating legal technology is sensible, as it ensures the future-proofing of regulations towards technological advancement. It is also sensible as it avoids the need to hermetically delineate what falls within “legal technology” and what does not. Such a technology-neutral approach has also been seen in existing legislation, such as the Electronic Transactions Act and the PDPA. The incrementalism that this approach offers also avoids confusion and conflicts (with existing regulations) that could arise in developing a new and broad regulation altogether. Nevertheless, it still remains essential that legal technology players are involved when it comes to the crafting of amendments to existing regulations, as such players provide the required specific technical and business knowledge to ensure that regulations support, rather than suppress, the use of legal technology.
Having explored the various issues that policymakers would consider tackling with any legal technology regulation, and having also explored the possible modalities through which such regulations should be introduced, we observe how other jurisdictions have introduced regulations to regulate legal technology. Rather than advocating for any particular position, the scan below is provided to share lessons from each jurisdiction that could be relevant to Singapore.
In France, legal technology is regulated by the Ethics Charter for an Online Legal Market and its Actors (the “Charter“). The Charter contains a broad definition of “Legal Tech actors”, referring to organisations that use technology to “develop, offer or provide products or services related to law” or allowing law end-users access to these products and services. The definition is kept broad to embrace a broad range of actors who apply technology in respect of legal services. Notably, the Charter is not limited to legal professionals. Article of the Charter 1 broadly refers to “an actor of Legal Tech” on top of legal professionals. The Charter aims for people to operate in the market with products and services that respect regulations. Security and confidentiality in Art 5 have the same principles as the Data Protection Regulation, as well as the concept of conflict of interest that is important to the legal industry. There are also remedies in place against those who breach those guidelines.
It is noted the French government has also introduced a regulation that disallows analytics on judicial decisions. As the first of its kind, this regulation was designed to address concerns of confidentiality and judges’ divergence from Civil Law norms.
Overall, it appears that the French regulatory framework appears to be take a broad approach, yet still generally encouraging towards legal technology providers.
Malaysia has arguably adopted a more restrictive regulatory framework. The authors of the Malaysian chapter in the State of Legal Innovation in the Asia-Pacific (“SOLIA”) 2020 Report posited that Malaysia’s less welcoming disposition towards innovation may be because of its precedent-based legal system. In recent years, the Malaysian Bar Council blocked three legal technology start-ups from providing their services in Malaysia. The Malaysian Legal Profession Act 1976 (“LPA(M)”) prohibits unauthorised persons from directly or indirectly carrying out acts of an advocate and solicitor. The Bar Council’s reading of the LPA(M) is much stricter than that in Singapore. The reasons given for the strict interpretation of LPA are as follows: First, to protect legal profession from unscrupulous vendors and/or access to products and services that do not comply with the LPA and related legislation. Second, to ensure that access to legal services is properly regulated so the general public is protected in terms of their access, including with regard to the quality of legal services. Further, the Legal Profession (Practice and Etiquette) Rules 1978 (“PER”) prohibits touting and division of costs or profits between legal practitioners and unqualified persons. The Malaysian Bar Council has deemed some services to be in breach of the LPA(M) or PER. The Future in Technology Committee has suggested some amendments to the LPA(M). Proposed changes may include allowing online submissions for practicing certificates, reports, compensation fund and more. On the whole, it appears that the Malaysian Bar Council has taken a more conservative approach towards the adoption of legal technology. The authors of the Malaysian chapter in the SOLIA Report 2020 opined that having a wide definition of legal technology may hinder the development of legal technology in Malaysia. While such an approach may come with good intentions (to protect the general public and the profession from the potential risks of technology), it is posited that regulators should avoid taking an approach that could hinder innovation and productivity.
As far as this author is aware, Germany adopts a restrictive regulatory framework. This can be seen by how, for instance, a Paper from the Federal Ministry of Justice stated that service portals with digital legal services should only be run by lawyers. Such a position would disallow service portals run by non-lawyers, thus limiting the development of legal technology to lawyers. Although the Paper further mentions that the ban on third-party ownership should be relaxed, no further details were canvassed. The Paper by Conference of Ministers of Justice discusses the “permission of contingency fees” but does not address the relaxation of the ban on third-party ownership.
The United Kingdom
In comparison to other countries, the UK’s regulatory framework is more liberal. The UK government has launched a regulatory sandbox which allows legal technology companies to test out their solutions with other stakeholders, including regulators. This enables the government to make findings specific to their country and tailor the regulations based on the insights they gleaned.
The Legal Services Board, which oversees the regulation of legal services in England and Wales, aims to build a regulatory climate that promotes innovation and enhances access to legal services. As a separate regulatory body to ensure that regulations for legal technology will protect public interest, it is likely that the UK will be forward-looking when implementing regulations. For instance, the Legal Services Board has also recommended taking active steps in regulation instead of a more reactionary approach. Such an approach would be beneficial in attempting to develop regulations that promote legal technology while reducing likelihood of undesirable outcomes. This would likely be helpful in Singapore as it can help to promote greater adoption of legal tech.
A way forward
Overall, it appears that existing legal services regulations in Singapore, such as the LPA and PCR, remain well-placed to address most, issues that may arise from the use of legal technology. As the analysis above shows, however, there are grey areas in certain aspects, such as the differences between legal information and legal advice arising from the use of legal chat-bots, and the distinction’s impact on the use of legal technology tools, that may call for legal technology-specific regulatory clarifications.
Hence, as a matter of first principles, it appears that the most efficacious way to regulate legal technology would be to take an “incremental” approach: adjusting the LPA and PCR to cover these specific issues in relation to legal technology. As legal tech services are tools that assist in the provision of legal services, the objectives that policy makers would want to achieve through regulating legal tech would be largely similar to regulating the legal profession. However, one key point to note is that the regulations concerning legal tech will have to be prepared to address any unique, and sometimes, unanticipated outcomes. Thus an incremental approach is recommended at this stage so that an appropriate standard for regulation can be determined.
Further, in light of growing role that legal technology professionals will play in the legal industry, it may also be more sensible that regulations be expanded to include legal technology service providers, as opposed to having to introduce entirely new regulations. To this end, it is suggested that since the amendments would greatly concern the legal technology service providers, having legal technology service providers’ inputs would be useful in discussions on any amendments to the LPA as their perspective is expected to offer balancing considerations from an innovation viewpoint.
 Mark D. Fenwick, Wulf A. Kaal Ph.D., “Regulation Tomorrow: What Happens When Technology Is Faster than the Law?” (“Fenwick, Kaal”), American University Business Law Review Volume 6, Issue 3, at p 572.
 Federation of Law Societies of Canada, Robots and Rule-Makers: New Frontiers for Legal Regulation: Report from the 2018 Annual Conference Charlottetown PEI October 17-18, 2018, online: https://flsc.ca/wp-content/uploads/2019/07/PEIReportERF.pdf.
 Fenwick, Kaal, supra n 1, at p 567.
 Tim Kirkman, Fiona Kirkman, “Legal Technology is Changing the Legal Profession. What Does this Mean for You?”, Law Gazette (November 2018), online: https://lawgazette.com.sg/practice/tech-talk/legal-technology-is-changing-the-legal-profession-what-does-this-mean-for-you/.
 Alvin Chen, Stella Chen, “Disruptive Legal Technologies – Is Ethics Catching Up?”, Law Gazette (August 2018), online: https://lawgazette.com.sg/feature/disruptive-legal-technologies-is-ethics-catching-up/.
 “Legal Information vs. Legal Advice: What is the difference?”, Centre for Public Legal Education Alberta (n.d.), online: https://www.cplea.ca/legal-information-vs-legal-advice-difference/.
 Robin Mills, “Reference service vs. Legal advice: Is it possible to draw the line”, 72 Law Libr. J. 179 (1979).
 While we were unable to find an exhaustive survey of all existing legal tech chatbots, a quick perusal of an online list of legal tech chatbots showed that the majority utilised decision-trees. Zane Ice, “The Playing Field of Legal Chatbots”, Medium, online: https://medium.com/legal-design-and-innovation/the-playing-field-of-legal-chatbots-58f2843ee9f4.
 Legal Profession Act (Chap 161), Legal Profession (Professional Conduct) Rules 2015 (“LPA”).
 Greg Henry, “What’s the difference between legal information and legal advice?”, Turtons (26 November 2019), online: https://www.turtons.com/blog/whats-the-difference-between-legal-information-and-legal-advice.
 Quoine Pte Ltd v B2C2 Ltd  SGCA(I) 02 at -.
 Esso Petroleum Co Ltd v Mardon  QB 801.
 LPA, supra n 9.
 Asia Law Network, https://www.asialawnetwork.com.
 SingaporeLegalAdvice.com, https://singaporelegaladvice.com.
 LPA, supra n 9.
 LPA, supra n 9. See also Jeffrey Pinsler, Legal Profession (Professional Conduct) Rules 2015: a commentary, at p 666-667.
 LPA, supra n 9, r 44(1)(a). See also Jeffrey Pinsler, Legal Profession (Professional Conduct) Rules 2015: a commentary, at p 669.
 Ibid. See also LPA, supra n 9, r 4.
 LPA, supra n 9.
 LPA, supra n 9, r 6.
 Personal Data Protection Act 2012 (No. 26 of 2012) (“PDPA”).
 Gabriele Buchholtz, “Artificial Intelligence and Legal Tech: Challenges to the Rule of Law”, Regulating Artificial Intelligence pp 175-198.
 France Bans Judge Analytics, 5 Years In Prison For Rule Breakers, artificiallawyer (4 June 2019), online: https://www.artificiallawyer.com/2019/06/04/france-bans-judge-analytics-5-years-in-prison-for-rule-breakers/.
 Fatimah Zahirah Mohd Damanhuri, “Malaysia – Embracing Technology In The Legal Industry — ‘Adapt’ Or ‘be Dropped’.”, Conventus Law (3 April 2019), online: http://www.conventuslaw.com/report/malaysia-embracing-technology-in-the-legal/.
 State of Legal Innovation in the Asia-Pacific 2020 (n.d.) (“SOLIA Report 2020”), online: https://drive.google.com/file/d/1QLiSwQ7XAkBIpXS55D4udr-fKMmfJarM/view at p 77.
 Kuek Ser Kwang Zhe, “New Horizons: Innovation and the legal profession”, The Edge Markets (14 July 2018), online: https://www.theedgemarkets.com/article/new-horizons-innovation-and-legal-profession.
 Malaysian Bar Council’s scrutiny of Dragon Law continues legal innovation debate (15 June 2016), online: https://themalaysianlawyer.com/2016/06/15/malaysian-bar-dragon-law-legal-innovation/.
 Zahirah, supra n 30.
 Kuek, supra n 32.
 June Low, “Legal technology Scene in Malaysia”, Legal Business World (19 February 2019), online:https://www.legalbusinessworld.com/post/2019/02/19/legal-tech-scene-in-malaysia.
 SOLIA Report 2020, supra n 31, at p 86.
 Earlybird Venture Capital, “Legal technology in Germany: A Thriving Industry at Risk of Over-Regulation”, online: https://medium.com/birds-view/legal-tech-in-germany-a-thriving-industry-at-risk-of-over-regulation-aeee1cda9a93.
 Earlybird Venture Capital, supra n 43.
 Stephen Mayson, UCL Centre for Ethics and Law, Independent Review of Legal Services Regulation: Assessment of the Current Regulatory Framework, Working Paper LSR-0 (March 2019), online: https://www.ucl.ac.uk/ethics-law/sites/ethics-law/files/irlsr_wp_lsr-0_assessment_1903_v2.pdf.
 “Plans unveiled for legal technology sandbox to fuel innovation in the UK”, The Global Legal Post (21 May 2020), online: https://www.globallegalpost.com/big-stories/plans-unveiled-for-legal-tech-sandbox-to-fuel-innovation-in-the-uk-87690498/; “LawTech UK”, Tech Nation, online: https://technation.io/lawtechuk-vision/#the-lawtech-sandbox.
 The Legal Services Board: Home, online: https://www.legalservicesboard.org.uk/about-us/who-we-are.
 Legal Services Board, Perspectives on Lawtech and Regulation: How can legal services regulation support responsible technological innovation that improves access to justice (April 2020), online: https://www.legalservicesboard.org.uk/wp-content/uploads/2020/04/LSB-Technical-Perspectives-FINAL.pdf.