11/14/25 Tech Tips
- Zoe Davis
- Nov 13
- 2 min read
What questions to ask before using AI as a lawyer?
Client Confidentiality
Not every AI platform covers attorney-client privilege as much as it should if it is to be used for law purposes. ChatGPT, for example, openly admits that the information is logged and open to subpoena. The information must be safeguarded as much as it would be with a human processor. Encryption, limits on data retention, and enterprise-grade compliance certifications are all necessary.
Is this Reasonable?
When asking an AI questions about legal documents or legal analysis, you should inspect where it's getting its reasoning from. Most major platforms are transparent about where they get their information from, but be sure to fact-check their sources. Do they reference credible materials (e.g., laws, research papers, official guidelines) to back up their claims? Make sure that if the analysis is based on a document (like a contract or regulation), it highlights the specific parts that influenced the outcome. They should also keep a record of how inputs were processed so users can trace the reasoning. You should also check its training data. Basically, is this AI just surfing the web for information? Is the information outdated as well? Laws change fast, and if the AI were based on out-of-date information, then it might give incorrect suggestions. If the AI is unable to be transparent or has unreliable training data, it is probably not a credible source to use and could leave you liable.
AI is Wrong Sometimes
No matter what AI tool or system you use at one point, it will get it wrong. Be sure to a) check all the work to make sure it doesn't look dubious, and b) the system you are working with should have some form of error reporting. Human review should be built into the system, and any contracts with vendors should outline liability terms. If it includes a verification and validation steps before results are given, and then are cross-checked against other legal sources, it greatly decreases the mistakes the AI could make.
I hope you have found these tech tips helpful! Please feel free to share these tips out or contact us with questions or comments.









Comments