Professional Conduct Rules That Apply to AI Assisted Work

The Law Society’s Code of Conduct for Solicitors sets out principles that directly apply to AI assisted work. Understanding these principles is essential because they define your legal and professional obligations.

Principle 6 requires that you uphold the rule of law and the proper administration of justice. This principle is engaged whenever you use AI in any work that touches the court system or that affects legal rights. If you use AI in a way that undermines the accuracy of your advice or that misleads a court, you breach Principle 6. The principle does not prohibit AI but requires that you use it in a way that maintains the integrity of the legal system.

Principle 2 requires that you act with integrity. This principle is directly relevant to AI use. If you present an AI-drafted document as your own analysis without disclosing the AI assistance, you may be acting without integrity. If you verify that an AI citation is fabricated and use it anyway, you breach integrity. Integrity means being honest about how the work was done and ensuring that the work is accurate.

Principle 5 requires that you provide a proper standard of service. This includes using appropriate tools and having proper expertise. Using AI without understanding its limitations or without verifying its output falls below the standard. A lawyer who delegates substantive legal analysis to an AI tool without checking the result is not providing a proper standard of service. The tool can assist but you remain responsible for the quality and accuracy of advice.

The Key Principles in the Code of Conduct also require that you have systems to manage risk. This includes systems to manage AI-related risk. If you permit junior staff to use AI tools without supervision or verification, you have failed in this duty. If you use AI without a clear protocol for checking output, you have failed in this duty. Risk management means having processes in place before you use the technology.

Rule 3.4 specifically addresses confidentiality and privilege. It requires that you keep client information confidential and that you do not waive privilege without good reason. This rule directly constrains your use of cloud-based AI tools. As discussed in an earlier lesson, disclosure of client information to an AI vendor may breach this rule if the client did not consent. Disclosure of privileged material may waive privilege.

Rule 2.1 requires that you have competence and knowledge to carry out your work. This includes competence to understand the AI tools you use and their limitations. If you use an AI language model for legal research without understanding that such models can fabricate citations, you lack competence. You must educate yourself about how the tools work and what risks they create.

The Bar Standards Board’s Core Duty 2 requires that barristers “do not mislead anyone, including the Court”. This duty applies directly to AI use. If you rely on AI-generated citations without verification and present them to the court, you are creating a risk that the court will be misled. The duty does not permit that risk.

Some practitioners ask whether they have “permission” from the Law Society or Bar Standards Board to use AI. The answer is that both bodies permit AI use but within strict conditions. Those conditions are the principles of conduct mentioned here. You have permission to use AI provided that you maintain integrity, provide proper service, protect confidentiality, do not waive privilege, and do not mislead the court.

A practical approach is to think of AI as a tool like any other. A lawyer can use a word processor to draft a document. But if the word processor contained errors and the lawyer did not check them, the lawyer remains liable. The word processor is just a tool. The lawyer’s judgment and verification remain essential. The same is true of AI. The tool can assist but does not replace your responsibility.