Accuracy

An AI language processing tool is trained form information inputted into it.  This means that when the algorithm was created, the decisions made by the creators about where the training information is sourced is critical to the output of information a user receives.  Some AI tools are not currently capable of providing of the moment up to date answers. As the legal industry is constantly changing with new case laws and legislation, there is the risk that AI tools will return inaccurate or biased information.   

In addition, a user has no way to know when an AI tool provides completely false information. An AI tool will always return an answer even if there is no answer to the question posed by the user. This means that if it does not find an answer to the question, it will return a wrong answer instead. The wrong answer is not flagged as incorrect or missing crucial context – the only way a user can guarantee the accuracy of any AI proclamation is by verifying that information themselves.  

 Furthermore, deep learning models—like the one used in ChatGPT’s underlying algorithm structure, do not allow the user to draw a thorough understanding of how a model comes to a prediction. This makes any answer ChatGPT provides difficult to interpret and harder to trust. The lawyer must always consult and confirm with other sources, verify the authenticity of the information provided and that it is up to date.  

AI is not a Legal Expert

Lawyers are often asked for advice that is unique or have no precedent and requires creative problem solving that factor in past experience from real life situations, combined with legal expertise.  AI is not a legal expert. It does not have the legal capacity or level of specialisation to interpret or analyse legal matters that may be specific in the context of that business transaction or legal proceeding.  

 The interpretation of legal documents is not a simple task.  It requires taking into consideration legal principles, context, intent of the parties, case law, nature of the transaction, and the business among multiple other factors, and thus extensive legal training.   

 AI does not resolve nuanced situations as it does not have the capability to perform human reasoning. It can only form an answer on the statistical patterns retrieved from the information that it has been trained on.   

Confidentiality

Chat GPT is not confidential and does not delete specific prompts from a user’s history. Open AI’s privacy policy states that the company collects personal information included in “input, file uploads, or feedback” that its users provide. Every conversation is logged, including any personal data shared, and it is then used as training data.   

The prompts inputted into the system are stored. These prompts are then fed into the system and returned as answers to other users prompts. This potentially breaches client lawyer confidentiality, as any information inputted into the chatbox has the potential to be shared with other users, and so has the potential to breach the lawyers professional code of conduct.  

Training and Development

Traditional training of new lawyers in law firms involves young associates learning their craft by carrying out routine tasks such as checking document reviews for accuracy. What happens when technology automates many of those routine tasks?  

The implementation of AI tools does have the potential to make legal work more efficient and accurate – but it may also make it harder for young lawyers to learn and develop necessary skills.   

Trainees may not get the experience of identifying and extracting contract terms and learning from their mistakes. A lawyer’s work is highly specialised and requires many years of training and accumulated experience to deliver results and client satisfaction. There is a very real concern that implementing AI into legal practice could affect the quality of the future legal workforce.   Having a basic understanding of how AI chatbots work is crucial for incorporating the technology into the legal practice.   

 

Here at LS Law, we do believe AI has a place, but in substantive legal advice and support. We are a team of industry legal experts, who have many years of experience in developing our legal knowledge and experience in the life science industry.

In our view, there is no substitute for the human context and legal training.