Artificial Intelligence and the legal system: Have we got the balance right?

Artificial Intelligence and the legal system: Have we got the balance right? 

    

- Parth Shah

The world we are living in is everchanging; continuous developments and innovations are shaping the modern world, with the paramount discovery being in ‘Artificial Intelligence.’ It is a problem solver. The legal profession is not exempt from its impact. Many would believe that it would be a disadvantage not to implement AI in a firm. But is this really so? The use of AI can open many possibilities: not only would it improve the access to justice, but also can bring in a greater clientele by increasing the efficiency of work. After all, time is what the firm is selling. Whilst on one level this could benefit both clients and the firm, it could result in job displacement, vulnerabilities in data protection and ethical risks, all which carries immense costs. Therefore, a firm looking to implement AI within their infrastructure would need to weigh up both the pros and cons, deciding how best to utilise AI in the short and long term. 

Income: A legal firm is a business. Like any business, income is what allows the firm to function - from paying wages and training future lawyers to hiring offices to run the firm. Therefore, the clients and cash flow the firms bring is of utmost importance. A firm’s vast earnings come from the number of hours that are billed. How could the introduction of AI influence this flow of income? The increase of efficiency AI provides by saving a significant amount of time would reduce the number of hours lawyers and paralegals can bill. Since the traditional method was to bill clients using the ‘hourly rate,’ which allowed for flexibility and transparency, the use of AI would make this extremely uneconomical for firms. Clients would benefit extremely by paying less for a task that can be completed at a faster rate through AI than if the task was worked on by a lawyer. On top of this, the cost for the firm to implement AI in their work is substantial, ranging from ‘$5,000 to $500,000’. This would suggest AI may disadvantage firms in both the short and long term. However, impending changes the firms are implementing would reverse these inconveniences. Restructuring the firm’s charging model to a ‘fixed fee’ model would result in AI improving the firm. This would mean that the firm has a guaranteed bill, meaning that the amount of time and resources spent on the case would not impact the income coming in. This would result in firms putting more money in AI to improve their efficiency, expediting tasks to handle a greater number of clients and increasing their income. However, many would argue that firms' quality of work and productivity would reduce, prioritising speed over quality. But this argument is rendered meaningless due to the profits the firm makes through client retention. Studies show that client retention by 5% can increase profitability by ‘25% to 95%’ because of the lower cost in retaining clients than acquiring others. Therefore, firms will also prioritise quality and results, making AI a tool that not only improves speed of these outcomes for the firm, but also benefits the client immensely in the work provided for them.

Whilst this may be theoretical, the use of AI to benefit both the client and the firm can be seen in a real-life case study. This involved the UK ‘Serious Fraud Office (SFO)’ investigating the famous ‘Rolls Royce’ for large-scale international bribery in 2017. This case involved the reviewing of 30 million + documents, which would have been impossible in time and cost if done manually for the SFO. Therefore, SFO worked in partnership with the AI company - ‘RAVN’ - to process these documents. This processed around 600,000 documents daily, cutting 80% of the workload in review. Ultimately, due to the inconsistency in which the company’s AI discovered, ‘Rolls Royce’ settled for £671 million, which would not have been possible without this AI tool. These staggering figures reflect the power of AI and the benefits were immense. Not only did this bring justice, it helped the SFO to secure future government funding and public trust, as well as the AI company gaining international recognition. This boosted credibility and influence across the entire legal field. Therefore, despite this case being in the public sector, this case reflects that AI can have significant benefits for all parties involved financially as well in seeking justice. But this case did not show that AI only has benefits. It highlighted that it could increase the number of unemployed lawyers. When the SFO partnered with the AI company, many barristers were offloaded because of how effective the AI was. Whilst the firm may benefit from the profit, employees may be disadvantaged. Not only the job insecurity, but it will create unequal access; those most comfortable with technology will benefit. However, this can also be balanced with the new career paths it opens as well as using the skills of workers to the greatest use. Despite there being ethical concerns of discrimination in terms of skills, it could also be balanced with the new skills it can offer to present and future employees. 

Data Protection:

Like most ideas and innovations, benefits always come with problems. Whilst AI allows firms to process and review ‘2000 times’ faster than human lawyers, the problems that arise are because of the type of data that is being handled with. Much of the information on these documents are extremely sensitive and a liability. It is an obligation for any lawyer to maintain confidentiality with clients' information. When AI is used to process this information, they become a third-party, which weakens confidentiality. However, poor handling of this information can have immense repercussions - if the AI breaches privacy, under the ‘UK GDPR and Data Protection Act’ (2018), firms can face huge fines, can face negligence claims as well as trust. This can reduce the number of clients retained, which can have huge losses for the firm. Furthermore, the greater firms push towards using AI, it opens the firm to vulnerabilities against cyber-attacks. This problem has already arisen in the past concerning ‘Orrick, Herrington & Sutcliffe LLP (USA).’ In April 2024, this law firm had to make an $8 million settlement over data breach affecting 600,000 individuals due to a malicious attack. Not only these costs, but extra costs also occur through the implementation of cybersecurity and staff training. All this would suggest AI can have disastrous outcomes for any law firm. So, is it worth this risk?

Use of AI in international law:

International law is the ‘set of rules and agreement that governs how countries interact with one another’. The main purpose of international law is to maintain peace, protect human rights and advocate security. These rules come from treaties e.g. the Paris Agreement, judicial decisions and unwritten rules that act in good faith. Therefore, the introduction of AI within this sector of law would have global implications. AI can assist in drafting international agreements as well as predicting outcomes of cross-border disputes - all which could improve how effective the law is globally. However, problems can occur within the International Humanitarian Law (IHL). These laws are used to regulate human behaviour, but if an autonomous weapon is used (which could cause mass destruction), who is at fault? The programmer, the state? These problems of ‘accountability’ are blurred with the use of AI; however, many still believe human oversight is still fundamental in tackling this problem. Furthermore, AI systems are being increasingly used in border control and surveillance, for example the AI biometric recognition software installed in Heathrow. However, under international human right laws, there are legal obligations to maintain privacy and to prevent discrimination, which is difficult when AI makes biased decisions. However, the ‘Council of Europe’s 2024 AI convention’ passed a treaty that forced governments to ensure AI operates within the rule of law in terms of human rights. The demand for transparency means that AI can be used fairly and still greatly impact international law enforcement. This is explored further through organisations such as the United Nations (UN) and G7 for a shared responsibility of AI to ensure fairness and encourage innovation with the use of AI internationally. In the future, AI can be used to arrive at agreements much more quickly than treaties and enforce much more effectively, however, this must not be at the expense of global stability and human rights.

Human vs AI:

The debate between the skills of humanity and the skills of artificial intelligence continue as AI enters the legal field. AI comes with many benefits for firms during cases: analysing thousands of documents to find inconsistencies and loopholes to be exploited. It could also offer equality in punishments and remove bias upon decision making. But can these machines ever truly understand justice when it is all about emotion and empathy? Especially during family and criminal law, logic and reason are not the only determining factors in the case. Rather empathy, genuine emotion and moral nuances in the case play a significant role - which AI can easily miss, but not by human lawyers, juries and judges. This is extremely evident in the Re G children [2006] UKHL case (names omitted for privacy protection). This case involved the splitting of a same-sex couple and the question argued was on the topic of who cares for the children. Lower courts ruled that the biologically linked mother should hold custody. But the case was re-opened in the House of Lords, in which the judge looked at the emotional reality of the case rather than biological facts. The judge stressed that the decision must be for the welfare of the children, ruling for shared custody. This decision would be illogical for AI to understand, which would suggest the need for humanity in the field of law. AI can only achieve justice to an extent. Rather, human lawyers are still needed to achieve justice when AI misses.

The Future in law:

By weighing all the pros and cons AI brings to the legal industry, making a ‘right’ decision is extremely difficult. Looking through any perspective, whether it is the lawyer, firm, client - the introduction of AI carries immense risks, but can also reap immense rewards. A solution to this problem could be to mediate halfway - implement AI slowly, allowing for current and future lawyers to adapt, aiding to reduce discrimination as well as reduce reliance on AI. With universities and firms already providing technological education, future lawyers will be able to use their empathy as well as technological knowledge to best support their clients and firm. Therefore, there will not be a conflict between the human and the AI - rather there will be a mutual understanding which will benefit all. 

1. Practical Law. Legal services fee models: an overview [Internet]. London: Thomson Reuters; [cited 2025 Nov 1]. Available from: https://uk.practicallaw.thomsonreuters.com/w-020-6926?contextData=(sc.Default)&transitionType=Default

2. LawWare. Law firm client retention: give ’em enough rope… [Internet]. Edinburgh: LawWare; [cited 2025 Nov 1]. Available from: https://www.lawware.co.uk/blog/law-firm-client-retention/

3. Future Processing. AI pricing: how much does Artificial Intelligence cost? [Internet]. Gliwice: Future Processing; 2025 Oct 22 [cited 2025 Nov 1]. Available from: https://www.future-processing.com/blog/ai-pricing-is-ai-expensive/

4. Sky News. Serious Fraud Office hires ‘artificial intelligence lawyer’ [Internet]. London: Sky News; 2018 Apr 11 [cited 2025 Nov 1]. Available from: https://news.sky.com/story/serious-fraud-office-hires-ai-lawyer-11325218

5. Legal Futures. Bad news for barristers: SFO adopts AI-powered document review after successful test in Rolls-Royce case[Internet]. London: Legal Futures; 2018 Apr 11 [cited 2025 Nov 1]. Available from: https://www.legalfutures.co.uk/latest-news/bad-news-for-barristers-sfo-adopts-ai-powered-document-review-after-successful-test-in-rolls-royce-case

6. Reuters. Law firm Orrick agrees to $8 mln settlement over breach client data [Internet]. London: Reuters; 2024 Apr 11 [cited 2025 Nov 1]. Available from: https://www.reuters.com/legal/litigation/law-firm-orrick-agrees-8-mln-settlement-over-breach-client-data-2024-04-11/

7. UK Parliament. In re G (Children) (FC) [2006] UKHL 43 [Internet]. London: House of Lords; 2006 Jul 26 [cited 2025 Nov 1]. Available from: https://publications.parliament.uk/pa/ld200506/ldjudgmt/jd060726/child-1.htm

8. King’s College London. The Praxis-Belief Continuum: Revealing customary international law through AI-assisted analysis [Internet]. London: King’s College London; 2025 Jun 5 [cited 2025 Nov 1]. Available from: https://www.kcl.ac.uk/the-praxis-belief-continuum-revealing-customary-international-law-through-ai-assisted-analysis

9. American Society of International Law (ASIL). Insights, Vol. 29, Issue 1 [Internet]. Washington DC: ASIL; [cited 2025 Nov 1]. Available from: https://www.asil.org/insights/volume/29/issue/1

Council of Europe. Council of Europe opens first-ever global treaty on AI for signature [Internet]. Strasbourg: Council of Europe; 2024 Sep 5 [cited 2025 Nov 1]. Available from: https://www.coe.int/en/web/portal/-/council-of-europe-opens-first-ever-global-treaty-on-ai-for-signature