What AI Chatbots Tell People Who Ask About a Criminal Charge
When someone is charged with a crime, or a family member is, one of the first things they may now do is type the situation into an AI chatbot. "I got a DUI last night in Texas, what happens now?" or "Can police search my car if I refuse?" The answer arrives in seconds, in confident paragraphs, at an hour when no law office is answering the phone.
This note looks at what those answers tend to look like and where they often fall short. It is meant for defense lawyers who want to understand what a prospective client may already believe before the first call. It is educational, it is not a study with measured results, and the tools change often, so run a few of your own questions to see what they say today.
The answers usually start general
Ask a chatbot about a criminal charge without naming a state, and most tools respond with a general outline: what the charge usually means, the typical stages of a case, possible penalties described in ranges, and a suggestion to talk to a lawyer. Many add a short disclaimer that they cannot give legal advice.
When a state is named, the better tools try to tailor the answer. Some get the broad strokes right. Problems show up in the details: a penalty range from an older version of the statute, a deadline that applies in a neighboring state, or a diversion program described as if it were available everywhere.
A person reading this at 1 a.m. has no way to tell which parts are reliable. That matters for how you write your own content and how you handle the first conversation.
Where the information comes from
Tools differ in how they source answers. Some rely mostly on what they learned during training, which can be many months or years out of date. Others search the web and show citations, so you can see which pages they drew from. Perplexity puts citations up front. ChatGPT and Gemini show sources when they search, and Google's AI Overviews link to pages under the summary.
When sources appear, they are often a mix of state government pages, legal information sites, large directories and individual law firm articles. A clear, accurate firm article on a specific charge in a specific state can be one of those sources. A vague page that says "contact us to learn your options" gives the tool nothing to cite.
For a closer look at how assistants pick which firms to mention when someone asks for a lawyer, see how AI assistants decide which defense lawyer to name.
Common gaps worth knowing about
Some patterns come up often enough to plan around:
- Local practice is missing. The tools rarely know how a particular county court handles arraignments or which local diversion options exist.
- Separate administrative processes get skipped. DUI answers often focus on the criminal case and mention the license hearing briefly or not at all.
- Timing is fuzzy. Chatbots tend to describe deadlines in general terms, or not at all, even where the real deadline is short.
- Immigration and professional license consequences get a sentence at most, though they can matter more to the client than the sentence itself.
These gaps are where a defense firm's own content can be most useful, and where the first consultation often needs to correct what the client read.
What clients may believe when they call
A caller who has read an AI answer may arrive with a clearer picture of the process, which can help. They may also arrive with a wrong idea held firmly. Someone told by a chatbot that a first offense "usually results in probation" may be surprised by the actual exposure in their county. Someone told they "should not speak to police" may already have done so and feel they ruined their case.
Some defense lawyers now ask early in intake whether the person has looked anything up, including with AI tools. It opens the door to correct misunderstandings without making the person feel foolish for searching.
Writing content that holds up
If AI tools are going to summarize your pages, write pages that survive being summarized. In practice that means:
- State the jurisdiction clearly on every charge page.
- Put the most important facts in plain sentences near the top, such as the deadline to request a license hearing or the general penalty range.
- Date your content and review it whenever the law changes.
- Explain local process from your own experience, which general sources cannot do.
Accuracy matters more than volume here. A page with a wrong deadline can be repeated by a chatbot to many readers. Check every statute and deadline against the current official source before you publish, and check your state bar's advertising rules on how you describe outcomes.
Questions about AI chatbots and criminal charges
Do chatbots tell people not to hire a lawyer?
The major tools generally suggest consulting a criminal defense lawyer, especially for anything beyond a minor infraction. The quality of the surrounding information varies a lot from tool to tool and question to question.
Should a defense firm warn visitors about AI answers?
A short, neutral note on a charge page that general information online, including from AI tools, may not reflect your state's current law or local practice can be helpful. Keep it factual.
Can a firm correct what a chatbot says about a charge?
Not directly. The practical route is publishing accurate, clearly written content and keeping your firm's details consistent across the web, so tools that search have better material to draw on.
More notes on how AI tools affect legal marketing are collected in our ChatGPT and AI search guides for lawyers.
SBD Marketing works only with criminal defense law firms, one firm per city. We help defense practices with websites, SEO for the charges they defend, visibility in AI search tools such as ChatGPT, Gemini, Claude and Perplexity, and SEO audits.