You are currently viewing Lawyers : How to Get Away with AI

In May 2023, a news item made the headlines in technology and legal news: “Because of ChatGPT, an American lawyer cites rulings… that never existed”, “An American lawyer used ChatGPT to prepare a case… and only cited fake rulings”, “My God, I regret it’: The lawyer who used ChatGPT says he was duped by AI”.

Indeed, an American lawyer used ChatGPT to prepare a case but cited rulings (court judgments) that never existed. This incident led to a fine of approximately €4,600 for the lawyer and his firm, highlighting the risks of using artificial intelligence (AI) in the judicial field. Let’s review what artificial intelligence is, the risks lawyers face when using it, and propose recommendations for better use of this technology.

Artificial intelligence in brief

Artificial intelligence (AI) is defined as the ability of machines to perform tasks that traditionally require human intelligence. There are two main types of AI: predictive AI and generative AI.

  • Predictive AI: Uses algorithms to analyze data and make predictions based on historical information.
  • Generative AI: Creates new content in response to requests (prompts), such as texts, images, or even videos. ChatGPT, Gemini, and Perplexity are examples of this technology.

Risks and disadvantages of generative AI for lawyers

The use of generative AI presents several risks for lawyers:

  • Bias: AI models can reproduce biases present in the training data, which can lead to discriminatory results.
  • Hallucinations: Models such as ChatGPT can generate incorrect or non-existent information due to the limitations of their training databases and algorithms. For example, ChatGPT can create fictitious court judgments, as in the case of the American lawyer, or provide false legal references.
  • Security and confidentiality risks: Queries entered into generative AI are often reused to automatically improve their training bases, which can lead to the use and subsequent disclosure of customers’ personal data “without their knowledge” to others.

These issues can result in professional liability for the lawyer, firstly if the information provided by the AI they used is incorrect, and secondly if the use of AI leads to negative consequences for their client or violates “attorney-client privilege“.

Advantages and best practices for using AI

Despite these risks, AI offers many advantages for lawyers, particularly in legal research and drafting. AI can analyze large amounts of existing legal data to provide relevant information quickly. AI can automatically generate drafts of legal documents, improve their consistency and accuracy, and integrate specific clauses based on needs.

Here are some AI best practice recommendations:

  • Verification of responses: Always verify the information provided by AI to correct any potential errors and hallucinations.
  • Anonymization of data: Do not enter personal data into AI systems or anonymize it to protect the confidentiality of client information.
  • Ongoing training: Lawyers need ongoing training in new technologies and AI to use these tools effectively and ethically.

As noted in a report by Thomson Reuters, “It is important for law firm leaders to understand that the evolution of AI will happen with or without them“. By adopting these practices and adapting to new technologies, lawyers will be able to reap the benefits of AI while minimizing risks. The mistakes made by the American lawyer demonstrate the importance of judicious and verified use of AI. With proper training and adapted practices, lawyers can “get away with AI“.


M2 Digital Economy Law – Class of 2023/2024

Sources : 

  1. À cause de ChatGPT, un avocat américain cite des arrêts… qui n’ont jamais existé
  2. Un avocat américain a utilisé ChatGPT pour préparer un procès… et n’a cité que des faux arrêts
  3. “Mon Dieu, je regrette “: L’avocat ayant utilisé ChatGPT estime s’être fait avoir par l’IA
  4. The new world is here for the legal industry… are you ready or not?
  5. Impact de l’intelligence artificielle sur les métiers du droit : Si l’IA l’avait fait, est-ce qu’on s’ferait l’effet que l’on se fait ?
  6. Le droit des affaires à l’ère de l’IA : Implications juridiques de l’intelligence artificielle dans les affaires

A propos de Freddy BASILA BULAMBO

En formation en master 2, droit de l'économie numérique à l'Université de Strasbourg.

Laisser un commentaire

Ce site utilise Akismet pour réduire les indésirables. En savoir plus sur comment les données de vos commentaires sont utilisées.