As reported by WINK News in Lee County, a Lee County woman was arrested after statements she allegedly typed into an AI chatbot were reported to law enforcement as threats against the Lee County Sheriff’s Office. According to the report, the statements were escalated through the AI company’s safety monitoring to a human review team, which then reported them to law enforcement. She reportedly faces a Florida charge of making a written threat of violence.

It remains to be seen whether she will be convicted. And a person is presumed innocent until proven guilty. An arrest report tells one side of the story at the very start of a case, and nothing here should be read as a comment on the facts or on her. (For more on how these investigations work, see our page on AI-related crimes in Tampa and across Florida.)

One Case, and the Start of a Trend

From where I sit as a criminal defense lawyer, this news is another example of what is certainly going to become a wave in the criminal justice system: people use chatbots and AI tools, and what they typed is later used against them in a legal proceeding. AI chat evidence won’t stay limited to threat cases. It can turn up in a fraud investigation, a custody fight, a theft case, or almost anywhere words typed to a machine are later read by people. (I wrote about this risk recently in Your AI Prompts Can Be Evidence.)

Every Third Party Is a Potential False Friend

In law school, we used to call any third party a potential “false friend.” You might tell someone something you believe is in confidence, and that person can turn around and repeat it to someone else. If what you said is a crime, or evidence of one, that someone else may be law enforcement.

The same is true of third-party businesses and software providers: Google, cloud storage companies, and artificial intelligence or superintelligence companies. The bottom line is that anything you say can and potentially will be used against you.

Platforms have rules. When someone uses an AI tool in a way that appears to violate that platform’s terms of service, the company may take action. When it looks like a serious potential criminal offense, monitoring should be expected. These companies are under a great deal of pressure to keep their networks free of crime and to make sure they are not participating in criminal activity themselves.

So it should come as no surprise that prompts, questions, and instructions given to an AI platform can result in information passing from the company to law enforcement as part of an investigation. According to WINK’s report, that is the path the Lee County statements reportedly took.

Why a Chatbot Feels Like No One Is Listening

The WINK report touched on this, too. People interviewed for the story noted that a conversation with AI can feel private and nonjudgmental. It can feel like writing in a private journal, or like talking to no one at all.

But AI chat evidence can exist in more than one place. Depending on the company, the product, and the user’s settings, conversations may be stored on the company’s servers, reviewed under its safety and usage policies, and preserved and later obtained by investigators through legal process. Copies may also sit on the phone or computer where the conversation took place. (Our national site explains the legal process side: How Police Get Your AI Chats, Images, and Account Records. Florida has already seen technology used to review what people say in other settings, such as jail calls; see Prisons Use AI to Monitor and Analyze Calls.)

Talk to a Lawyer, Not a Chatbot

There is no attorney-client privilege with a chatbot. If you are worried about an investigation, something you said, or something you did, a chatbot is not the place to work it out. A lawyer is. Conversations with your own attorney are confidential in a way an AI chat is not. (More on that in AI Is Not Your Lawyer.)

That is not a reason to fear technology, which has real value. It is a reason to understand that AI chat evidence is real, and that what you type may not stay between you and the screen.

Questions About AI Chats and Your Case? Let’s Talk.

If you would like to speak with an experienced criminal defense lawyer about how AI chats or other digital evidence could affect an investigation or case in Florida, or you have further questions, don’t hesitate to contact Attorney Ron Frey of The Frey Law Firm, LLC. Your consultation is free and confidential. You can talk with us by phone, by video, or at our Tampa office, and when the situation calls for it, we can come to you.

Call or text (813) 775-2276. Rapid response, 24/7.

This post is for general information only. It is not legal advice, and reading it or contacting us does not create an attorney-client relationship.

Laws involving AI and digital evidence are changing quickly and vary from state to state, and how they apply depends on the facts of each case and on how courts interpret them. Please don’t rely on this post, or on any other information online, to make decisions about your situation. Talk with a lawyer.

Florida and federal law both address these situations directly. Florida has laws covering sexual images involving minors, including images that are AI-generated or altered, and laws covering sexual deepfakes of real people made or shared without their consent. Federal law can apply as well. (For an overview, see our page on AI-related crimes in Tampa.)

This post explains, in general terms, where the law stands in Florida and what the process tends to look like, so you have a clearer picture of the questions to bring to a lawyer. For a broader look at how these cases work locally, see internet crime charges in Tampa.

Where the Law Generally Stands in Florida

A few points hold true across most of these cases:

• They are treated as serious criminal matters. Florida law addresses AI-generated and altered sexual images involving minors, and sexual deepfakes of adults made or shared without consent, as crimes.
• More than one kind of conduct can be involved. Depending on the law and the facts, creating, requesting, possessing, or sharing material can each be at issue.
• Federal law can apply too. Some investigations that begin with local police end up in federal court, and some begin with federal agents. (See our pages on federal criminal defense and sex offense defense.)
• The laws are relatively new and have been updated recently. Older information online may no longer be accurate.

The consequences of a conviction can be serious and long-lasting, affecting a person’s freedom, family, career, and reputation. Exactly what applies in a given case depends on the facts, and that’s a conversation for a lawyer.

How These Investigations Usually Unfold

AI-image and deepfake investigations tend to follow a familiar path:

• A starting point. That might be a report from a technology platform, a complaint from the person depicted, a tip, or information from an unrelated investigation.
• Records and devices. Investigators may request account records from companies and seek search warrants for phones and computers. Forensic examinations can take months, so a search without an arrest doesn’t necessarily mean the matter is over.
• Contact. At some point, investigators often want to talk with the person they are looking at, sometimes casually and sometimes at the door during a search.
• A charging decision. Prosecutors decide whether to file charges, and whether the case belongs in state or federal court.

(For more on how investigators gather digital evidence, see How Police Get Your AI Chats, Images, and Account Records and What Should You Do if the Police or FBI Search Your Home? on our national site, InternetCrime.com.)

What the Government Has to Prove

In every criminal case, the person accused is presumed innocent, and the government has to prove its case beyond a reasonable doubt. That applies to cases involving technology the same way it applies to any other case.

In digital cases, the questions that often come up include:

• Who actually created, requested, possessed, or shared the material?
• What does the digital evidence show, and what doesn’t it show?
• How was the evidence collected, and did investigators stay within the warrant?
• Were the person’s rights respected during the search and any questioning?

A lawyer can look at those questions in light of the actual facts and explain what they mean for your situation.

Talking With Investigators

Statements made to investigators can become evidence. Conversations that happen under stress, without the full picture, can be easy to misunderstand. A lawyer can communicate with investigators for you and help you understand where things stand before any decisions are made.

“It’s much better to get your questions answered before decisions are made than after.”
— Attorney Ron Frey

Attorney Ron Frey’s Take

Attorney Ron Frey has over two decades of experience defending people in state and federal criminal cases. In his experience, the early stage of an investigation is often when a lawyer’s involvement matters most, and a confidential consultation is the place to get real answers about your own situation.

“If you’re wondering whether you should talk to a lawyer, that’s often a very strong sign that you need one.”
— Attorney Ron Frey

(For more on that point, see Should I Hire an Attorney?)

Questions to Bring to a Lawyer

People in this situation often ask questions like these:

• What exactly am I being accused of, and by whom?
• Investigators took my devices but didn’t arrest me. What happens now?
• How should I respond if a detective or agent contacts me?
• Is this a state case, a federal case, or could it become either?
• What could this mean for my job or my professional license?

Each of those questions turns on the facts of your case, and a lawyer can answer them confidentially.

Accused or Under Investigation? Let’s Talk.

If you would like to speak with an experienced criminal defense lawyer about an investigation or charge involving AI images, AI videos, or deepfakes, or you have further questions, don’t hesitate to contact Attorney Ron Frey of The Frey Law Firm, LLC. He represents people in Tampa Bay and throughout Florida. Your consultation is free and confidential. You can talk with us by phone, by video, or at our Tampa office, and when the situation calls for it, we can come to you.

Call or text (813) 775-2276. Rapid response, 24/7.

This post is for general information only. It is not legal advice, and reading it or contacting us does not create an attorney-client relationship.

Laws involving artificial intelligence are new and changing quickly, and how they apply depends on the facts of each case and on how courts interpret them. Please don’t rely on this post, or on any other information online, to make decisions about your situation. Talk with a lawyer.

It feels private. It feels like talking to an advisor. It isn’t either one.

Generative AI tools can explain general legal concepts, and lots of people find that helpful. But a chatbot is not your attorney, and the law doesn’t treat your conversations with it like conversations with your lawyer. (If you have questions about an AI-related investigation, see our page on AI-related crimes in Tampa.)

“ChatGPT, Grok, Claude: none of them is your lawyer. They don’t owe you loyalty, they don’t owe you confidentiality, and they can’t stand between you and a subpoena.”

Attorney Ron Frey

What the Attorney-Client Privilege Actually Protects

The attorney-client privilege protects confidential communications between a client and a lawyer made for the purpose of getting legal advice. It exists so people can tell their lawyer the whole truth without fearing that those words will be turned against them. On top of that, lawyers owe their clients a professional duty of confidentiality, and they answer to the courts and the Bar if they break it.

An artificial intelligence platform has none of that, and neither will the more autonomous AI agents now arriving. It isn’t a lawyer, it isn’t licensed, and it owes you no professional duties. And depending on the company’s terms and privacy policy, it may keep what you type, use it to improve its products, and disclose it to others, including in response to legal process.

Courts Are Starting to Draw the Line

Recent cases have made the same basic point: chats with an AI tool about your own legal situation are not treated like conversations with your attorney. Courts have rejected claims that those chats are protected by the attorney-client privilege, and they have treated a chatbot as what it is: not a lawyer, and not a confidential advisor.

The law is still developing, and results can differ depending on the facts and the court. What courts have not done is treat a chatbot as a substitute for a lawyer. If you’re under investigation or worried about possible charges, confiding in an AI tool about your situation is a real risk.

“When you talk to your lawyer, the law is built to protect that conversation. When you talk to a chatbot about your case, you may be writing the government’s next exhibit.”

Attorney Ron Frey

Subpoenas, Discovery, and Search Warrants

Even apart from privilege, AI chat records can be reached in several ways:

• Search warrants for phones and computers can capture saved chats, exported documents, screenshots, and synced data.
• Legal process to the provider. Federal law lets the government seek account records from online service providers and, with a warrant, the contents of communications.
• Civil discovery and preservation orders. In civil lawsuits, courts can require companies to preserve AI chat records, and opposing parties can seek relevant chats through discovery. Recent cases have also treated conversations people voluntarily send to an AI platform as having limited privacy protection.

In practice, anything you type into an AI tool about your situation may someday be read by a prosecutor, an opposing lawyer, or a judge.

Talking to Your Own Lawyer Is Different

When you speak with a defense lawyer, even in a first consultation before you hire anyone, the conversation is generally confidential. Your lawyer can hear the whole story, including the parts you’re most worried about, and give advice that fits your facts, your jurisdiction, and your goals. If a lawyer decides that technology, including AI, would help in your defense, it can be used under that lawyer’s direction and with that protection in mind.

“AI can be a useful tool. It’s not a substitute for a lawyer who is obligated to keep your secrets and fight for you.”

Attorney Ron Frey

Have Questions You’ve Been Asking a Chatbot? Ask a Lawyer Instead.

If you’ve been using AI to try to make sense of an investigation, a charge, or something you’re worried about, Attorney Ron Frey is glad to talk with you. Your consultation is free and confidential. You can talk with us by phone, by video, or at our Tampa office, and when the situation calls for it, we can come to you.

Call or text (813) 775-2276. Rapid response, 24/7.

This post is for general information only. It is not legal advice, and reading it or contacting us does not create an attorney-client relationship.

Generative AI chatbots have taken that to another level. More people now ask ChatGPT, Grok, Claude, Gemini, and other AI tools the questions they used to type into Google, and they ask them in full sentences, with details, often late at night when they’re worried. Those prompts, the AI’s answers, and any images the tool creates can be evidence too, in criminal cases and in civil lawsuits. (For an overview of AI-related criminal charges in Florida, see our page on AI-related crimes in Tampa.)

Lately, Attorney Ron Frey has been hearing from a growing number of people who are worried about something they typed into an AI tool. This post looks at why that concern is real, and why context matters so much.

“A Google search shows what you were curious about. An AI chat can read like a diary, and prosecutors read diaries.”

Attorney Ron Frey

Why AI Prompts Are Different From Search Terms

A search is usually a few words. An AI conversation is a back-and-forth. People explain their situation, describe what happened, ask follow-up questions, and sometimes ask outright whether something they did, or are thinking about doing, is illegal. Image tools add another layer: the prompt that produced an image, the image itself, and the timestamps can all be saved.

That makes AI chats potentially powerful evidence of:

• Intent and planning: what someone was trying to accomplish, and when.
• Knowledge: what someone understood about a situation or a risk.
• State of mind: worry, anger, or a fixation on a subject.
• Timeline: exactly when a question was asked or an image was made.

“People type things into an AI that they would never say out loud to another human being. That’s exactly why those words can end up in a courtroom.”

Attorney Ron Frey

This Is Already Happening

AI chats are already appearing in court filings. In recent criminal cases, prosecutors have pointed to questions people typed into AI tools, and to images those tools generated, as part of the story they tell a jury. Those are allegations that have to be tested like any other evidence. Separately, recent cases have rejected claims that chats with an AI tool about a person’s own case are protected by the attorney-client privilege. (We cover that point in AI Is Not Your Lawyer.)

Civil Cases, Too

This isn’t only a criminal-law issue. In divorces, employment disputes, business fights, and other civil lawsuits, the other side can ask for relevant information through discovery, and AI chat histories can fall within those requests. How long companies keep records, and who can get them, depends on the company, the account, and the court process involved.

How AI Records Are Collected

In general terms, investigators have two places to look:

• Your devices. A search warrant for a phone or computer can reach app data, saved images, downloads, screenshots, and anything synced from the cloud.
• The company’s records. Under federal law, the government can seek account information from online service providers, and with a warrant it can seek the contents of communications. Providers can also be asked to preserve records while investigators obtain legal process.

How much a company keeps, and for how long, depends on the company, the account settings, and sometimes court orders. Investigators may be able to get it without your cooperation.

AI Is Not Your Lawyer

When you talk with an attorney about your legal situation, those communications are generally protected by the attorney-client privilege, subject to the rules of confidentiality. That protection is one of the reasons people call a lawyer when they’re worried.

A conversation with a chatbot, a robot, or any other artificial intelligence tool is different. There is no attorney-client privilege with AI. If those records are subject to a search warrant or a subpoena, the conversation can not only be discovered and revealed, it can potentially be used against a person in a court of law.

“If you have questions about the law or about your conduct, call a lawyer. Don’t ask a chatbot.”

Attorney Ron Frey

(For more on this point, see AI Is Not Your Lawyer.)

A Prompt Is Not a Confession

None of this means a prompt proves anything. People ask artificial intelligence tools questions out of curiosity, anxiety, research, fiction writing, or because they’re trying to understand a news story. A question about whether something is illegal is often the opposite of a plan to do it. Someone else may have been using the account or the device. And a few lines pulled out of a long conversation can mean something very different once you read the whole thing.

“A prompt is not a confession. Who typed it, why they typed it, and what the whole conversation actually says, that’s where the defense begins.”

Attorney Ron Frey

Worried About Something You Typed? Let’s Talk.

If you have questions about the law, about something you typed into an AI tool, or about conduct that could come up in an investigation, call a lawyer. Attorney Ron Frey is glad to talk it through with you confidentially. Your initial consultation is free. You can talk with us by phone, by video, or at our Tampa office, and when the situation calls for it, we can come to you.

Call or text (813) 775-2276. Rapid response, 24/7.

This post is for general information only. It is not legal advice, and reading it or contacting us does not create an attorney-client relationship.

Notice to all Participants to the Communication

All monitored calls shall contain a prompt which clearly identifies the call as coming from a Florida Department of Corrections institution and advise that the call is subject to being monitored and recorded by the Department of Corrections. Likewise, signs are required to be posted on or near each phone which state that calls are subject to being monitored or recorded.

Additionally, it is standard practice for defense counsel to advise incarcerated clients, and their families, that communications are recorded and monitored. I advise clients to imagine law enforcement, the prosecutor, judge, and jury are all listening in on their jail calls. Despite warnings, that imagined scenario still manages to manifest itself in certain cases. When jail recordings are introduced as evidence, they typically include discussion of criminal activity, even candid confessions.  

Artificial Intelligence

As reported in several media outlets, earlier this year the Florida Department of Corrections contracted with LeoTech, a company that produced a patented artificial intelligence software platform named Verus. According to the LeoTech website, Verus provides continuous real-time monitoring of communications and enables users to search “through millions of transcripts at sub-second speeds.” The A.I. software provides for searches by keyword, customizable alerts, the ability to investigate and uncover links and connections amongst the data collected, and even the ability to identify slang and coded lexical categories, terms, and synonyms. 

Needle in a Haystack

Although it has been standard practice to record jail calls, the ability to effectively examine the content has been limited by the labor and costs required. Widespread examination would be equivalent to looking for a needle in a haystack. Effective mass-investigation of the communications of 80,000 inmates would not economical or feasible without A.I. It has been reported that the contract with LeoTech enables prisons to record and scan up to 50 million minutes of conversations. What used to be impractical, if not impossible, is now at the fingertips of prison officials.

Implications

The use of A.I. will assist prison officials and law enforcement by enhancing the efficiency of call monitoring. However, it raises concerns from privacy advocates, particularly regarding the potential for abuse or misuse of the technology, as well as the potential issues pertaining to data access and storage duration. For the defense practitioner, it is now more important than ever to advise clients, and their families, that calls while incarcerated are not simply subject to recording and monitoring, but also to real-time analysis and investigation conducted by artificial intelligence software.

Author: Ron Frey – The Frey Law Firm, LLC