ChatGPT and Internet Crime Investigations

ChatGPT is the most widely used AI chatbot in the world. People use it for work, school, and personal questions, and many treat it like a private notebook or even a confidant. ChatGPT conversations are not legally privileged the way conversations with a lawyer can be, and they have begun to appear in criminal investigations. This page gives general background on ChatGPT and how investigations involving it often happen.

Background and History of ChatGPT

ChatGPT was released by OpenAI in November 2022. OpenAI was founded in 2015 as a nonprofit research lab and later created a for-profit arm. ChatGPT reached a mass audience faster than almost any consumer app before it, and OpenAI has since added image generation, voice, file uploads, web search, memory, and agent features.

ChatGPT Features and Services

ChatGPT lets users type or speak prompts and receive generated text, images, and other outputs. Users can upload photos and documents, create custom GPTs, share conversations by link, and turn on memory so ChatGPT remembers details across chats. Conversation history is generally saved to the account unless a user turns history off or uses a temporary chat, and OpenAI’s privacy settings and retention practices may change over time.

How People Use ChatGPT

ChatGPT is used legitimately by a very large number of people for writing, research, studying, brainstorming, coding, translation, and everyday questions. Businesses use ChatGPT’s business plans to draft documents, analyze data, and support customers, and many schools and universities make it available to students and staff. OpenAI also offers ChatGPT Gov for U.S. government agencies, and in January 2025 OpenAI said that more than 90,000 users across more than 3,500 federal, state, and local government agencies had used ChatGPT to support their work. For many people, ChatGPT is a powerful productivity tool at home, at school, and at work.

How Many People Use ChatGPT?

OpenAI said in September 2026 that ChatGPT has 1.2 billion weekly users. We did not locate an official U.S.-only figure.

How Investigations Involving ChatGPT Often Begin

Many investigations involving ChatGPT do not start with a police officer stumbling onto something. They start with a report. Online providers based in the United States are required by federal law to report apparent child sexual abuse material that they become aware of to the National Center for Missing & Exploited Children (NCMEC), which runs the CyberTipline. NCMEC reviews each report, tries to work out where the user is located, and makes reports available to the appropriate law enforcement agency, often an Internet Crimes Against Children (ICAC) task force or a federal agency such as Homeland Security Investigations or the FBI. For more on what happens after a report is made, see what a CyberTipline report is and what happens next.

AI platforms add something new to investigations: the user’s own words. Prompts, uploaded files, generated images, and conversation history can all be stored by the provider, and they can also be found on a phone or computer. Investigators may seek AI records to show what a person was thinking or planning, or to show that a person tried to create or upload illegal content. The firm discusses how police obtain these records in how police get AI records and how chats are used in court in AI chats and prompts as evidence.

NCMEC’s 2025 data lists 182,844 CyberTipline reports from OpenAI, many relating to images uploaded to or requested from its tools. Other ChatGPT evidence comes from phones and computers seized for a different reason, where investigators find conversations that they argue are relevant.

A general caution: do not create, request, or share sexual or intimate images of any real person without that person’s consent, and never of a minor. AI-generated images can lead to criminal investigations even when no real photograph was used. See AI-generated images and deepfakes.

What Records ChatGPT May Have and How Police Request Them

When investigators want records from ChatGPT, they generally have to use legal process. In broad terms, basic account and subscriber information is often requested with a subpoena, more detailed records with a court order, and the content of messages or files with a search warrant, although the exact rules depend on the type of data, the jurisdiction, and the circumstances. Investigators also commonly send a preservation request, which asks the company to hold onto existing records while the legal process is prepared. What a company actually has depends on its products, its retention practices, and the user’s own settings, and those practices change over time.

OpenAI publishes a law enforcement policy describing how it handles requests. In general terms, OpenAI may hold account information (name, email address, payment information for paid plans), IP addresses and log data, and, with a search warrant, conversation content and uploaded files, depending on the user’s settings and OpenAI’s retention at the time. OpenAI also accepts preservation requests and emergency requests.

Connecting a ChatGPT Account to a Person

A record showing that a ChatGPT account did something is not the same thing as proof that a particular person did it. Investigators typically try to connect an account to a person using things like the email address or phone number used to register, IP addresses from logins, subscriber records from an internet or wireless provider, device identifiers, payment records, and what is found on devices during a search. Each link in that chain can raise questions: shared homes and shared Wi-Fi, devices used by more than one person, VPNs, carrier networks that place many customers behind one IP address, compromised accounts, and gaps in the records. These are the kinds of issues a defense attorney may look at closely, case by case, under current law. The firm discusses this further in how police connect online accounts to a person.

Examples From Reported Cases

The examples below are drawn from public announcements and news reports. They are described generally and without names. Charges are only accusations, and anyone charged is presumed innocent unless proven guilty in court. These summaries are not a prediction of how any other case will turn out.

• In Vermont in 2025, OpenAI reported an uploaded image of suspected child sexual abuse material to NCMEC. Investigators tied the account to a man, searched his residence, and arrested him. The charges are allegations.
• In Missouri in 2025, a young man was charged with vandalizing cars after police cited a ChatGPT conversation found on his phone, along with phone location data. The charges are allegations.
• In California in 2025, federal prosecutors charging a man accused of starting the Palisades Fire cited an image and prompts he allegedly created with ChatGPT months before the fire. The charges are allegations.

Related Articles on InternetCrime.com

• Platform Investigations: all platforms
• AI Crimes
• AI Chats and Prompts as Evidence
• How Police Get AI Records
• AI Chatbot Prompts as Evidence
• AI-Generated Images and Deepfakes
• AI Evidence in Court
• What Is a CyberTipline Report?

If You Are Under Investigation Involving ChatGPT

If police have contacted you, if a search warrant has been served, or if you believe an investigation involving a ChatGPT account may be underway, it is important to get legal advice before taking any action on your own. Attorney Ron Frey generally advises speaking with an attorney before answering questions or agreeing to a search. Whether a person may be required to provide a password or unlock a device depends on the circumstances, the jurisdiction, and any search warrant or court order involved, so that decision should be made with an attorney’s advice (see should I provide my password to the police or FBI?).

You have the right to remain silent. Anyone accused of a crime is presumed innocent unless proven guilty in court. Laws change and vary from state to state and between state and federal court, and the way they apply depends on the specific facts. Please don’t rely on the internet or on AI tools to evaluate your situation. Every case is different, and contacting an attorney is a priority to get tailored advice.

Attorney Ron Frey of The Frey Law Firm, LLC has been practicing law for over two decades and represents people nationally in internet and computer crime investigations and charges. He is a conference speaker and published author.

Attorney Ron Frey Is Available for a Free Initial Confidential Consultation: Call or Text (844) 766-3739. In Florida, call (813) 775-2276. You can also reach the firm through the contact page.

References

• OpenAI, “DevDay 2026 recap” (Sept. 2026)
• NCMEC, “2025 CyberTipline Reports by Electronic Service Providers (ESP)”
• OpenAI, Law Enforcement Policy (Dec. 2025)
• Vermont Attorney General’s Office, press release (June 2025)
• Above the Law, on the Missouri ChatGPT case (Oct. 2025)
• BBC News, “ChatGPT image snares suspect in deadly Pacific Palisades fire” (Oct. 2025)
• OpenAI, “Introducing ChatGPT Gov” (Jan. 2025)
• Wikipedia, “ChatGPT” (background)

Important: This page provides general information only. It is not legal advice, and it is not a complete or current statement of the law or of any company’s policies. Laws, court decisions, and platform practices change, and their effect depends on the facts of each case and on how courts interpret them. User figures and report counts come from the sources listed and may since have changed. Platform names and logos are trademarks of their owners and are used only to identify the services discussed; The Frey Law Firm, LLC is not affiliated with or endorsed by any of these companies. Please don’t rely on this page, or on any other information online, to make decisions about your situation. Talk with a lawyer. Reading this page or contacting The Frey Law Firm, LLC does not create an attorney-client relationship. Last reviewed: October 26, 2026.