Legal tech – is AI rendering justice?
Advances in artificial intelligence (AI), in particular machine learning (ML), can transform legal practices. For instance, lawyers may utilise “legal tech” to select cases with highest chances of success.
Article information and share options
Legal tech is on the rise
Over the past decades, AI in law has advanced from academic research to impactful practical application. A 2020 OECD report lists numerous examples of AI that can be applied in the law profession,1 including supporting judges in courts to reduce backlogs in cases,2 increasing legal access to more citizens,3 and reducing uncertainty around the application of financial regulations.4
Several high-profile legal tech start-ups are valued at more than USD 1 billion, thus reaching “unicorn status”. 5 Simultaneously, many large and well-known law and consulting firms are investing in those legal tech start-ups that provide services to help lawyers in their daily tasks, from legal research to contract and litigation document generation.6
Challenges and limitations remain
Still, use of AI in law continues to face many constraints, not all of which are technical. Most AI systems rely on the right to access court data for their analysis. However, bulk access to this data in a comprehensive fashion remains difficult.7 On the technical front, key challenges persist, including automatic detection of the structure of cases (eg, identifying the conclusion), and encoding the logic of statutes.
Law firms typically use software developed by specialized software companies. Early adoption of inadequate or inefficient software may lead to incorrect or insufficient outputs. A further risk emerges from the question of whether biases of algorithms trained on biased datasets can be worse than the biases of humans.8 If biased software leads to incorrect advice, professional indemnity claims may emerge against lawyers, and liability claims may be lodged against the software companies.
How will AI impact the legal system?
AI could impact the law itself by fostering the idea of “legal singularity,” the notion that the need for interpretation can be removed from answers to legal questions.9 Some even raise the concern whether the law could lose its ability to evolve if, with the application of AI, it becomes mechanically based on prior arguments.10
Another risk pertains to auditing implementation requiring both legal and software engineering skills. Gatekeepers to the law (lawyers/judges/legislators) will typically not be experts in reading code, and software engineers will typically not be experts in interpreting law. This raises the question of transparency and monitoring the technology.
But one of the biggest direct impacts to insurers could stem from lawyers using legal tech systems to zero in on those lawsuits/cases with most chance of success. This could conflict with the legal purpose of providing justice for all, and also lead to a significant rise in claims loads.
Further Information
References
1 Mohun, J., and Roberts, A., “Cracking the Code: Rulemaking for humans and machines,” OECD Working Papers on Public Governance 42, 2020.
2 “Can AI Be a Fair Judge in Court? Estonia Thinks So,” Wired, 25 March 2019.
3 See in France Mes Aides, and for the issues that it created: Alauzen, M.,“Splendeurs et misère d‘une start-up d‘Etat: Dispute dans la lutte contre le non-recours aux droits sociaux en France (2013 – 2020),” Réseaux 225 (121 – 150), 2021.
4 See for instance the project in the UK led by the FCA, “Digital regulatory reporting” which seek to encode regulations (last consulted on Jan 17, 2021).
5 “The battle to win at legal tech,” Financial Times, 27 May 2021.
6 “The Future of Law Firms (and Lawyers) in the Age of Artificial Intelligence,”
American Bar Association, 2 October 2020.
7 The UK is aiming at publishing in a central form all court judgement only in April 2022.See The National Archives (2021). The National Archives to publish court and tribunal judgments, Jun 16 (last consulted on Jan 17, 2022).
8 See for instance: van den Hoven, E., “Hermeneutical injustice and the computational turn in law” Journal of Cross-Discipllinary Research in Computational Law, 1(1), 2022. However, with the literature on legal biases being so enormous, incl. on the software COMPAS used by US courts to determine chances of recidivism, we refrain from providing a comprehensive literature on the subject.
9 Markou, C., and Deakin, “Is Law Computable? Critical Perspectives on Law by
Artificial Intelligence,” Hart Publishing, 2020.
10 Hildebrandt, M. “Code-driven law: Scaling the past and freezing the future,”
Critical Perspectives on Law and Artificial Intelligence, 2020