top of page

AI could be a great tool for law firms but how practical is it ?

Updated: May 28

Despite the undeniable promise of AI in the legal field, the current environment presents significant obstacles that must be overcome. It is essential for law firms to understand the limitations and potential risks of AI technology. Only then will they be able to capitalize on the potential advantages and mitigate the potential disadvantages of incorporating AI into their professional practice.


An objective of less ‘Artificial’ and more 'Intelligence'

Artificial General Intelligence (AGI) has the potential to revolutionize the legal profession. This technology, which can perform tasks that would require human intelligence, could theoretically handle the work of a lawyer. However, we are still far from reaching this level of sophistication. The present reality is that we are still at the stage of utilizing narrow generative AI, which necessitates the expert skills of professionals who have an in-depth understanding of both AI and law to produce a coherent and relevant document for clients.

Firms' primary responsibilities consist of drafting and editing. Theoretically, generative AI could assist with both, but it is difficult to implement unless firms are willing to invest in developing appropriate tools and provide training to their lawyers on how to use them.

Scattered tools

A central challenge is the lack of comprehensive AI tools. Today's legal profession requires multiple distinct tasks, from drafting legal documents to interpreting complex legislation. Regrettably, there is no single AI tool capable of executing all of these tasks in one fell swoop. This limitation means that lawyers must often navigate a web of different tools, requiring both time and specialized knowledge to use effectively. For example, a lawyer seeking to review a complex agreement must employ different AI tools for different sections of the document, a task that demands both a broad skillset and a solid grasp of the relevant legal frameworks.

Power and limitations of AI tools

ChatGPT, developed by OpenAI, is one of the most renowned language models today. Capable of generating human-like text based on given prompts, it has found broad application across many sectors, including law. Its strength lies in its ability to understand and generate text contextually, making it a potentially invaluable tool for tasks like drafting legal documents or preparing briefs. However, the use of ChatGPT in a professional legal setting is not without limitations.

Firstly, token restrictions are a significant impediment. In the context of language models, a token can represent a character, a word, or anything in between. Every piece of text input or output from ChatGPT consumes a certain number of tokens, and there's a limit to how many tokens the model can handle in a single interaction. This means that extremely lengthy documents may need to be broken down into smaller parts for processing, leading to potential issues with continuity and context.

Version compatibility issues also pose challenges. Like any software, ChatGPT is subject to updates and revisions. Newer versions may include enhancements that increase its effectiveness or address previous limitations.

The necessity for specific plugins adds another layer of complexity. These plugins, which extend ChatGPT's core capabilities, can be incredibly useful, allowing the model to interface with different software or providing additional functionality. But they also require understanding and training to use effectively, further raising the bar for effective utilization of the tool.

Moreover, these limitations lead to a fragmented user experience. A lawyer might need to use ChatGPT for drafting a section of a contract, another AI tool for reviewing past legal cases, and yet another tool for analyzing the context of the transaction. This need to 'stitch together' outputs from different tools can lead to a discontinuous workflow and potentially error-prone outputs. The necessity to manually oversee and review each stage of the process also increases the lawyer's work, somewhat mitigating the efficiencies that AI is supposed to bring to the table.

Outdated information

Law is an inherently dynamic field. Legislation is regularly being updated, court precedents are frequently being set, and interpretations of legal texts can change swiftly. This fluidity ensures that the law keeps pace with societal changes and ethical considerations. However, it also means that any AI tool employed within this field must have access to the most current information in order to provide accurate and reliable assistance.

The challenge with AI tools like ChatGPT is that while they are incredibly powerful in processing and generating text based on their training data, they do not update their knowledge base in real time. This leads to a knowledge gap, where the AI's understanding of legal matters becomes outdated over time.

As a result, a legal professional using an AI tool could find themselves faced with an interpretation of a legal document that deviates from current legal understanding. For instance, if there have been recent amendments to a law or new case law developments that significantly alter the interpretation of certain legal provisions, these changes would not be reflected in the AI's output.

This knowledge gap poses several risks. For one, it can lead to misunderstandings and misinterpretations that could mislead clients and other legal practitioners. Relying on outdated legal advice could lead to incorrect decision-making, which could have serious legal repercussions for clients.

As a result, an AI's interpretation of a legal document could deviate from the current legal understanding, potentially misleading clients and damaging the lawyer's credibility.

Clients' confidentiality at stake

The incorporation of AI in legal practices brings along its own set of security risks that cannot be overlooked. These risks can significantly compromise the sanctity of confidentiality, a cornerstone of the legal profession. AI tools, such as ChatGPT, although transformational in their utility, can inadvertently threaten this principle if not managed appropriately.

One of the major concerns stems from how AI systems process and retain data. When information is fed into an AI system for processing, it's tokenized - effectively transformed into a format the AI can understand and learn from. There's a risk that this data, including sensitive and confidential information, could potentially be stored as part of the system's training data.

To further compound this issue, current AI systems do not have a mechanism to remove or 'forget' specific data once it has been learned. This characteristic resembles blockchain technology, in which data once recorded cannot be modified or deleted. The inability to 'unlearn' or erase specific data presents a clear risk of long-term data retention, which could be exploited if the system's security is breached.

Law firms that aren't sufficiently familiar with the intricacies of AI technology might inadvertently expose sensitive client data. This risk could result from a lack of awareness about the data handling practices of AI models or from a misunderstanding of how to securely integrate these tools into their workflow.

For instance, a lawyer could mistakenly use ChatGPT to process a document containing sensitive client information, potentially exposing that data. Likewise, law firms might unintentionally store confidential data on insecure servers or cloud systems while using AI tools, leaving them vulnerable to cyber-attacks.

These risks highlight the pressing need for law firms to establish robust data security protocols when implementing AI technology. Adequate training on the safe use of AI, coupled with stringent cybersecurity measures, is crucial to protect the confidentiality of client information and to maintain trust in the legal process. As AI continues to make entry into the legal field, law firms must remain vigilant about the unique security risks it presents.


Whereas ChatGPT and similar AI tools offer promising prospects for the legal industry, they currently serve more as a supplement than a comprehensive solution for legal professionals. For effective and accurate results, a solid understanding of both legal frameworks and AI technology is required due to the complexities of legal practise and the limitations of existing AI tools. Future legal practitioners can anticipate an increasingly integrated and effective experience as technology continues to advance.


bottom of page