Can I Use ChatGPT or Other AI During My California Family Law Case?

Nearly half of American adults have used AI chatbots, according to the June Pew Research Center report.  What once seemed like an unfamiliar technology has quickly become part of how people search for information, write messages, and make everyday decisions. It’s easy to see why.

Unlike a traditional Google or Bing search, where you might enter a few keywords and review several websites, AI lets you explain your situation, ask follow-up questions, and receive personalized answers almost immediately.

But that convenience can be risky when you are embroiled in a family law dispute, since chatbots are designed to keep you engaged.

You might begin by asking about California child custody laws. A few questions later, you are explaining what happened during your last custody exchange, what your spouse said, or why you disagree with your attorney’s advice.

It may feel like an ordinary conversation. Yet you could be creating a detailed record of your thoughts, concerns, and intentions that you never expected anyone else to read.

Consider this the family law version of the Miranda warning: Anything you tell a chatbot can and will be used against you in a court of law.

What Counts as AI During a California Family Law Case?

AI is not limited to the chatbot you open when you need an answer. It is increasingly built into the apps, search engines, and communication tools people use every day.

Some of the most common examples include:

• AI chatbots: ChatGPT, Claude, Gemini, Microsoft Copilot, Grok, Perplexity, Kimi, and DeepSeek.
• AI-powered search features: Google’s AI Mode and AI Overviews, and AI-powered search features in Bing.
• Social media and messaging assistants: Meta AI in Facebook, Instagram, Messenger, and WhatsApp, and Snapchat’s My AI.
• Email and writing tools: Gemini in Gmail, Copilot in Outlook and Word, and Grammarly’s AI writing features.
• Meeting and transcription assistants: Otter.ai, Fireflies.ai, Zoom AI Companion, and AI-powered phone transcription tools.

This is not an exhaustive list. New AI tools and features are rolled out regularly, including within apps and services people already use.

These tools do not all handle information in the same way. Some generate answers from questions you type, while others process documents, communications, or recordings. Their privacy settings and data retention practices can also differ.

Understanding those differences matters because you may be sharing information with an AI service without thinking about where that information goes afterward.

Is What You Type Into AI Private?

Not necessarily. Different AI companies have different policies governing what they collect, how long they retain information, and when they may disclose it.

Some save your conversation history to an account. Others offer temporary conversations or settings that limit how long they retain information or how they use it. Depending on the service, information may still be kept for a period after you delete it from your visible history.

Deleting a conversation from your screen does not mean the information has immediately disappeared from the company’s servers. Some data may remain temporarily or be retained for legal, security, or other reasons allowed by law.

Under certain circumstances, companies may be required to provide your user data in response to valid legal requests. 

That doesn’t mean that every AI company stores everything forever, or that anyone involved in your case can automatically access your search history. But it does mean that you should never assume an AI conversation is confidential.

Can AI Conversations Be Used Against You in California Family Court?

Yes, potentially. California has no specific law that automatically makes AI conversations admissible in court. But existing discovery and evidence laws may allow relevant AI chats to be obtained and used against you during a family law case.

One way the other party may obtain AI conversations is through discovery, the legal process through which parties request information from one another. Under California Code of Civil Procedure Section 2017.010, parties may request relevant, nonprivileged information, including electronically stored records.

Depending on what they contain, your AI conversations could fall within that category, just like emails or text messages. But obtaining a conversation is not the same as being allowed to use it as evidence in court.

Under California Evidence Code Section 1220, a person’s own statements may be admitted against them. That means questions, explanations, or comments you made to AI could potentially be presented by the opposing party.

Even then, the conversation must be relevant, authenticated, and otherwise admissible.

For example, the opposing party cannot automatically obtain your entire AI history simply because you are getting divorced. And even if they are allowed to read your AI chats, that doesn’t mean they can admit them in court. 

Importantly, what you tell AI is not necessarily the same as what AI tells you. An AI-generated answer is not automatically treated as proof that you intended to follow its suggestions.

The context of your conversation also matters. If the opposing party presents a misleading excerpt, your attorney may be able to challenge it or introduce relevant portions of the exchange to explain what you meant.

What U.S. v. Heppner Revealed About AI and Attorney-Client Privilege

In United States v. Heppner, No. 25 Cr. 503 (JSR), 2026 WL 436479 (S.D.N.Y. Feb. 17, 2026), a federal court considered whether conversations with a consumer AI chatbot could be protected in the same way as communications with an attorney.

Bradley Heppner, a former corporate executive facing a federal fraud investigation, had been using Anthropic’s Claude to prepare for possible criminal charges.

After learning that he was the target of the investigation, Heppner entered information into Claude, including information he had learned from his lawyers. He used the tool to develop reports about possible defense arguments, legal issues, and case strategy.

He subsequently shared the AI-generated materials with his attorneys. But when FBI agents searched his home in November 2025, they seized electronic devices containing documents reflecting those AI exchanges.

Heppner’s lawyers argued that the documents should be protected because he had created them to help prepare his defense and communicate with counsel. The court rejected that argument, finding that it was not protected either by attorney-client privilege or the work-product protection that applies to notes, memos, or research by a lawyer in a case.

Why Weren’t the Conversations Protected?

To understand the decision, it helps to distinguish two legal protections:

• Attorney-client privilege generally protects confidential communications between you and your attorney made to obtain or provide legal advice.
• The work-product doctrine shields an attorney’s legal thinking and certain materials prepared as part of their work on a case.

The judge found that Heppner’s AI exchanges did not meet those requirements.

Claude was not his attorney. The communications were made with a third-party platform whose privacy terms allowed certain data collection and disclosures. Although Heppner used the results to communicate with his lawyers, they had not directed him to create the materials.

Crucially, sharing the AI-generated materials with his attorneys afterward did not make them privileged. The ruling allowed prosecutors to review the seized documents, although it did not determine whether they could be admitted as evidence at trial.

Heppner was a federal criminal case in New York involving a consumer AI tool, not a California family law dispute. Some business and enterprise AI products offer stronger confidentiality protections, and using AI at an attorney’s direction may raise different legal considerations.

However, neither a paid subscription nor an attorney’s involvement automatically guarantees privilege. Whether you must disclose AI conversations depends on the circumstances and applicable legal protections.

The case highlights an important risk: information you believe you are privately preparing for your lawyer may not be legally protected. For that reason, we recommend avoiding AI for anything related to your ongoing family law case unless your attorney has specifically approved its use and advised you on what information you can share.

How Your AI Conversations Could Affect a Case

The concern is not limited to conversations in which you deliberately reveal legal strategy. Even the way you phrase an ordinary question can raise concerns when viewed outside its original context.

Consider a parent who asks AI how to stop the other parent from seeing the children. Perhaps the real concern involves a safety issue or repeated problems during exchanges. Without that context, however, the question could be presented as evidence of an intention to interfere with parenting time.

Or consider someone asking whether delaying a bonus, changing compensation, or transferring money could affect what they owe during divorce. Those kinds of questions could raise concerns about financial disclosure or efforts to reduce support obligations.

Neither example proves wrongdoing. A question is not necessarily an admission that someone acted on it.

But if the other side obtains those AI conversations, they may try to use them to:

• Challenge your credibility by comparing what you told AI with your testimony or court declarations
• Question your intentions concerning custody, parenting time, or communication with the other parent
• Raise concerns about financial disclosures, assets, income, or support
• Identify possible arguments or positions you were considering during the dispute

How much weight those statements carry depends on their wording, context, and connection to the issues being decided.

But you probably don’t want to be on the witness stand in court trying to explain what you meant by a late-night hypothetical you asked a chatbot.

What You Should Avoid Sharing With AI During Your Family Law Case

While your case is pending, avoid using consumer AI chatbots and built-in assistants to:

• Vent about your spouse, co-parent, children, or case. Emotional statements about the other parent or your children could be taken out of context and used to question your intentions or credibility.
• Seek case-specific legal advice or strategy. This includes questions about custody, support, property division, negotiations, settlement positions, or how to gain an advantage in court.
• Draft or revise communications about the dispute. Avoid asking AI to draft text messages, emails, or responses to your spouse or co-parent, especially on contested issues.
• Research how to win a case or respond to allegations in your case. Even the way you phrase a question may reveal concerns, intended arguments, or possible strategies.
• Upload confidential or sensitive documents. That includes communications with your attorney, financial records, custody evaluation reports, mediation materials, and documents containing information about your children.

Another reason to avoid relying on AI for your legal case: the answer may be wrong. AI programs have been known to hallucinate responses. A chatbot might invent an imaginary legal precedent, misunderstand California law, or present an exception to a rule as a rule.

In In re Domestic Partnership of Campos & Munoz (2026) 118 Cal.App.5th 1112, a California family law dispute over a pet dog involved fabricated case citations, some potentially generated by AI. The Court of Appeal sanctioned an attorney $5,000 for citing fictitious authorities in appellate filings and continuing to defend them after their authenticity was challenged.

For people representing themselves, the risk is relying on an apparently convincing answer that misstates the law or supports a court filing with information that has never been verified.

Using AI for Court Documents

If asking AI a question is risky, relying on it to create a court filing is downright dangerous. If you use AI to help prepare a legal document, make sure to follow any applicable court rules or orders, and never submit it without verifying all of the facts, legal citations, and quotations to ensure nothing has been hallucinated or misquoted.

All of this is covered by the California Courts’ guidance on AI, which warns that it can produce incorrect legal information, give advice that does not fit your situation, and should not be given private information.

Can You Still Use AI for Other Things?

Yes. Using AI for unrelated work, recipes, travel planning, or general everyday questions does not carry the same case-specific concerns. The important distinction is whether you are sharing information about your dispute or relying on AI to make decisions that could affect it.

For your own legal case, discuss those questions directly with your attorney.

What If You Have Already Shared Information About Your Case With AI?

If you have already discussed your issues with an AI tool, that does not automatically mean the information will be discovered or harm your case.

The first step is to stop adding further case-related information and speak with your attorney about what has already been shared.

Do not delete the conversations to make them disappear either. California courts may consider the deletion to be destruction of relevant evidence. You have a duty to preserve relevant records when you anticipate litigation, even before a case is formally filed.

Instead, take these precautions:

1. Preserve relevant information. Speak with your attorney about keeping AI conversations and whether any automatic-deletion settings need attention
2. Explain what you shared. Tell your attorney which tools you used and whether you entered legal advice, financial information, custody details, or other sensitive materials
3. Do not change the record. Avoid altering, fabricating, or reconstructing conversations to make earlier statements appear different
4. Seek advice before responding to requests. If the opposing party requests your AI records, have your attorney assess the request and any applicable objections or protections

Remember, using AI does not mean you have done anything improper. The goal is to address any potential concerns before they create further complications.

Protecting Your Interests in an AI-Driven World