AI-Assisted Legal Research: Enhancing or Undermining Judicial Independence?

by

 Andrei-Ioan Muresan

CyJurII Scholar

 on 11 July 2026 

Introduction:

AI legal research is a very important component within today’s digital world[1]. This is because it helps lawyers and other legal experts to quickly find relevant information, such as case laws, statutes, treaties, and many others[2]. In addition to this, according to a Thomson Reuters report, AI can save up to 240 hours of work per year from lawyers[3]. However, doing legal research with AI can also be dangerous. This is because it could mislead legal experts with false case number citation and many other things4. Therefore, this research paper will discuss the benefits, but also the danger of using AI tools for legal research. In addition, it will also discuss a case law that shows the dangers of over relying on AI tools.

The benefits of using AI tools for legal research:

AI tools are being used increasingly more frequently by legal experts in building up their own legal research. This turned out to have plenty of advantages to it, which will be discussed here. Firstly, AI provides legal research tools that can analyze large volumes of data quickly, specifically in minutes, instead of hours or maybe even days[4]. As a result, lawyers can focus more on strategic tasks related to their research[5]. Secondly, with the help of NLP and machine learning, AI can minimize human errors and deliver more context-aware and precise search results[6]. Thirdly, AI can quickly look through many previous cases and analyze important factors such as jurisdictions, case facts, defense strategies, etc[7]. Therefore, with these insights, lawyers can change their strategies, and thereby increase their client’s chances of a successful outcome[8]. Fourthly, AI helps law firms to stay updated while laws and judgments change frequently[9]. This means that the system continuously keeps an eye on regulations and statutes in case laws and thereby helps lawyers respond promptly to new legal developments, if needed11. Fifthly, through AI case law analysis, institutions can study past data, such as legal precedents, in order to predict future case outcomes and highlighting risks[10]. 

Finally, AI systems can find compliance risks, monitor regulations and flag issues that might be missed, automate regulatory tracking and enable proactive mitigation before issues go out of control[11]. However, AI legal tools can also cause other problems, if legal experts do not use them properly, which will be discussed in the next section.

The problems of using AI tools for legal research:

While AI tools for legal research provide many benefits, they can also cause a lot of problems, if not used carefully. One such problem relates to bias and discrimination[12]. What this means is that if an AI platform is trained on biased data, it will most likely lead to discriminatory outputs, which could then pose risks for lawyers using these types of technologies or even influence crucial aspects of the litigation process[13]. 

Another problem concerns data privacy and security[14]. What this means is that many AI developers rely on the information that users give for the training and improvement of their models[15]. However, lawyers still need to ensure the client-attorney privilege through not giving important information and through making sure that the platform has strong data privacy and security safeguards in place18. Lastly, the most dangerous problem in relation to legal research AI tools is the following, namely that these particular AI tools are susceptible to hallucinations[16]. What this means is that an AI tool can make a lawyer believe that they were provided with real case laws or case law citations, when in fact, that case law or case law citation has never existed in the first place[17]. Therefore, this signifies that lawyers and legal experts should also use their own existing knowledge and not automatically take what the AI tool gives them for granted21. Nonetheless, in the next section, there will be a discussion of a case law related to AI hallucinations and how these can catch lawyers off-guard.

Case law related to AI hallucinations:

A great example regarding AI hallucinations in law comes from the important South African case law known as The Northbound Processing (Pty) Ltd v South African Diamond and Precious Metals Regulator and Others [2025] ZAGPJHC 538[18]. Essentially, this particular case pitted the concept of AI liability with the concept of accountability and the doctrine of ‘’Duty to Know’’.

This particular saga initially began with a dispute between Northbound Processing and the Regulator, regarding a refining license[19]. However, while preparing to give the license in the final judgement, the judge accidentally discovered that Northbound Processing’s lawyers have used and referred themselves to fake case laws and citations[20]. These fake case laws were found out to have come from an AI tool known as ‘’Legal Genius’’, which was supposedly trained on South African legal judgments and legislation. This is therefore a classic example of AI hallucinations[21].

Conclusion:

To summarize, using AI tools for legal research is both beneficial and dangerous to legal experts. Essentially, these AI tools can help lawyers to look quickly through past case laws and help them draft a strategy, so as to help their own clients win their cases. Also, it can help them to find compliance risks, find issues that might have been missed and correct human errors. However, it can also be dangerous, as AI tools might steal people’s important data and or even give lawyers false case citations.

Bibliography:

Bloomberg Law, ‘’Can You Use AI for Legal Research?’’, 2024. Available at:

https://pro.bloomberglaw.com/insights/technology/can-you-use-ai-for-legal-research/ (Accessed on the 16th of June 2026).  

Griffith, Kelly, ‘’How do legal professionals effectively use AI for legal research?’’, 2025.

Available at: https://legal.thomsonreuters.com/blog/how-to-effectively-use-ai-for-legal-research/ (Accessed on the 16th of June 2026). 

A3Logics, ‘’AI for Legal Research: Key Features, Benefits and Future Trends’’, 2025. Available at: https://www.a3logics.com/blog/ai-for-legal-research/ (Accessed on the 16th of June 2026).  

Bloomberg Law, ‘’What Are the Risks of AI in Law Firms?’’, 2024. Available at:

https://pro.bloomberglaw.com/insights/technology/what-are-the-risks-of-ai-in-law-firms/ (Accessed on the 16th of June 2026).  

Osborn Wellsted Paulsen, ‘’It Aint My Fault - Accountability and the Duty to Know’’, 2025.

Available                                                                                                                   at:

https://www.owppartners.com/insights/the-northbound-case-accountability-and-the-duty-to-k now (Accessed on the 16th of June 2026). 

End notes

[1] Bloomberg Law, ‘’Can You Use AI for Legal Research?’’, 2024.

[2] Griffith, Kelly, ‘’How do legal professionals effectively use AI for legal research?’’, 2025.

[3] Ibid. 4 Ibid.

[4] A3Logics, ‘’AI for Legal Research: Key Features, Benefits and Future Trends’’, 2025.

[5] Ibid.

[6] Ibid.

[7] Ibid.

[8] Ibid.

[9] Ibid. 11 Ibid.

[10] Ibid.

[11] Ibid.

[12] Bloomberg Law, ‘’What Are the Risks of AI in Law Firms?’’, 2024.

[13] Ibid.

[14] Ibid.

[15] Ibid. 18 Ibid.

[16] Bloomberg Law, ‘’Can You Use AI for Legal Research?’’, 2024.

[17] Ibid. 21 Ibid.

[18] Osborn Wellsted Paulsen, ‘’It Aint My Fault - Accountability and the Duty to Know’’, 2025.

[19] Ibid. 

[20] Ibid.

[21] Ibid.