AI Crimes in Massachusetts: Deepfakes, AI Nudes, ChatGPT Chats and AI Note Takers

Artificial intelligence (AI) programs such as ChatGPT, Claude and Gemini write text and answer questions. The same programs make realistic images and video, imitate a person’s voice from a short recording, and transcribe meetings and phone calls. Since OpenAI released ChatGPT to the public in November 2022, these programs have become free or low-cost apps on ordinary phones and computers. This page explains the Massachusetts and federal crimes a person can be charged with for misusing AI programs.

Most people use AI programs for work, school and everyday questions. A person can also use the same programs to misuse another person’s face, voice or name. A person can generate a fake sexual image of a real person, impersonate someone on a phone call or in a text message, invent evidence such as a screenshot or a recording, or secretly record and transcribe a private conversation. A student or a professional with no criminal record can face a criminal charge for conduct that takes a few minutes on a phone.

A person in Massachusetts can be charged with criminal harassment for sharing an AI-generated nude image of a real person, with identity fraud for using a cloned voice to ask someone for money, or with a felony for secretly recording a conversation with an AI note-taking app. What the person typed into ChatGPT, Claude or another AI chatbot is not private. Police can obtain those chats with a search warrant. Prosecutors can then use the chats as evidence in court.

The sections below answer whether it is illegal to make or share AI nudes and deepfakes in Massachusetts, whether ChatGPT conversations can be used against you in court, and whether you can record a meeting or call with an AI note taker. The sections also cover AI voice clone scams, AI fake IDs, texts written by AI while a restraining order is in effect, and police use of facial recognition.

Massachusetts AI Laws

Massachusetts police and prosecutors charge AI-related conduct under the same criminal statutes that apply to conduct without AI. The criminal harassment statute, G.L. c. 265, § 43A, is the Massachusetts criminal statute that expressly names computer-generated images. The Legislature added that language in Chapter 118 of the Acts of 2024, which took effect on September 18, 2024.

As of September 2026, a second change is pending. In July 2026, the Massachusetts Senate passed an amendment that would make it a crime to use AI to create sexual images of children. The amendment is part of an economic development bill. The amendment becomes law only if the House and Senate conference committee keeps it in the final bill and the Governor signs the bill.

When the police do not arrest the person, the police usually apply to the court for a criminal complaint. A clerk-magistrate then decides whether to issue the complaint, often after a private clerk-magistrate hearing held before any arraignment.

AI Nudes and Deepfake Images of Adults

Sharing an AI nude image or sexual deepfake of a real, identifiable person can be charged as criminal harassment in Massachusetts under G.L. c. 265, § 43A(b)(2). A college student who uses a “nudify” app to turn a classmate’s photograph into a nude image and posts the image in a group chat can face that charge.

Section 43A(b)(2) punishes a person who knowingly distributes visual material showing another person who is nude, partially nude or engaged in sexual conduct. The statute provides that visual material includes material produced by digitization. The statute defines digitization as creating or altering visual material, including through computer-generated images, so that the material would falsely appear to a reasonable person to be an authentic image of the person depicted. The offense is the distribution. The Commonwealth must prove that the defendant sent, posted or otherwise shared the image.

The Commonwealth must prove that the distribution caused the person depicted physical injury, economic injury or substantial emotional distress. The Commonwealth must also prove that the defendant intended to harm, harass, intimidate, threaten, coerce or cause substantial emotional distress. In the alternative, the Commonwealth can prove that the defendant acted with reckless disregard for three things. The three things are the likelihood of harm to the person depicted, that person’s lack of consent to the distribution, and that person’s reasonable expectation that the image would remain private.

If a defendant is convicted under § 43A(b)(2), the judge can impose up to two and one half years in a house of correction, a fine of up to $10,000, or both. The District Court model jury instruction for this offense is Instruction 6.800, revised in February 2026.

Defense counsel examines account records, device data and metadata to determine whether the defendant created or sent the image. Defense counsel also examines whether a reasonable person would take the image for a real photograph. The statute’s definition of digitization requires an image that would falsely appear to a reasonable person to be authentic. The elements and defenses are set out in detail on the firm’s page on defending an AI deepfake image charge.

Federal law applies to the same conduct. The TAKE IT DOWN Act, signed on May 19, 2025, added 47 U.S.C. § 223(h). Federal prosecutors can charge a person who knowingly publishes, through an online service, an AI-made intimate image of an identifiable adult when the person intends harm or the publication causes harm.

If a defendant is convicted of that federal offense, the judge can impose up to two years in prison. For an AI-made intimate image of a minor, the maximum is three years. Threatening to publish an AI-made intimate image in order to coerce or extort someone is a separate federal crime.

An AI-made sexual image can also be the basis of a restraining order. Under G.L. c. 209A, § 1, publishing or attempting to publish sexually explicit images of a family or household member is a single act of coercive control when the act is intended to threaten, harass or control that person and causes that person to reasonably fear physical harm or have a reduced sense of physical safety or autonomy.

Threatening to publish sexually explicit images is part of the pattern definition of coercive control in the same section. A person depicted in an AI-made image can ask a judge for a 209A restraining order on that basis.

AI-Generated Child Sexual Abuse Material and Student Deepfakes

Federal law makes it a crime to produce, distribute or possess an image created, adapted or modified to appear that an identifiable minor is engaging in sexually explicit conduct. Federal prosecutors in Massachusetts can bring those charges today under 18 U.S.C. §§ 2252A and 2256(8)(C). In addition, 18 U.S.C. § 1466A punishes an obscene visual depiction of any kind showing a minor engaging in sexually explicit conduct, including a drawing, cartoon, sculpture or painting.

Massachusetts law on AI-generated child sexual abuse material is changing. The Senate amendment passed in July 2026 would make it a crime under state law to use AI to create sexual images of children.

A high school student who uses an app to make nude images of classmates from their school photographs can face a delinquency complaint in the Juvenile Court and a school discipline case at the same time. Chapter 118 of the Acts of 2024 also created G.L. c. 272, § 29D, a separate offense for a minor who possesses or distributes sexual images of other minors. The criminal case and the school case are described together on the firm’s page on AI deepfake charges against Massachusetts students.

AI Voice Clone Scams, Fake Accounts and Impersonation

Using an AI voice clone to pose as another person and ask for money can be charged as identity fraud in Massachusetts under G.L. c. 266, § 37E(b). A caller who uses an AI copy of a grandson’s voice to tell the grandson’s grandmother that he is in jail and needs bail money can face that charge.

Section 37E(b) punishes a person who, with intent to defraud, poses as another person without that person’s express authorization. The person must use the other person’s personal identifying information to obtain money or anything of value, or to harass someone. The statute defines personal identifying information to include a name, address, telephone number and similar identifiers. A caller who gives the grandson’s name on the call uses the grandson’s name, which is personal identifying information under the statute.

If a defendant is convicted under § 37E(b), the judge can impose up to two and one half years in a house of correction, a fine of up to $5,000, or both. When the grandmother pays, the Commonwealth can also charge larceny by false pretenses under G.L. c. 266, § 30.

A person who uses AI to create fake social media accounts and floods a former partner with messages can be charged with criminal harassment under § 43A(a). The Commonwealth must prove a knowing pattern of conduct or series of acts on at least three separate occasions, directed at a specific person. The Commonwealth must also prove that the conduct seriously alarmed that person and would cause a reasonable person substantial emotional distress. Finally, the Commonwealth must prove that the defendant acted willfully and maliciously.

In Commonwealth v. Johnson, 470 Mass. 300 (2014), the Supreme Judicial Court upheld criminal harassment convictions based on false online advertisements that listed the victims’ address and telephone number and brought strangers to their home. When the messages include a threat, the Commonwealth can charge stalking under G.L. c. 265, § 43, which is a felony. Criminal harassment charges based on online speech are subject to First Amendment limits.

Texts Written by ChatGPT and Restraining Order Violations

A defendant who sends a text message written by ChatGPT or another chatbot to the person protected by a restraining order violates the order’s no-contact provision. A defendant under a 209A abuse prevention order who asks a chatbot to write an apology to a former partner, and then texts the apology, has contacted the former partner. The Commonwealth can also charge a violation when an app the defendant set up sends a message to the protected person automatically.

A violation of a 209A restraining order is a crime under G.L. c. 209A, § 7. If a defendant is convicted, the judge can impose up to two and one half years in a house of correction, a fine of up to $5,000, or both. A violation of a 258E harassment prevention order is a separate crime under G.L. c. 258E, § 9. Answers to the questions defendants ask most often about contact with a protected person are in the firm’s 209A and 258E violation FAQs.

The person who asks for a restraining order can also face charges over fake evidence made with AI. A plaintiff who files a 209A complaint or a 258E complaint signs an affidavit under the penalties of perjury. Under G.L. c. 268, § 1A, a person who signs a written statement under the penalties of perjury commits perjury if the statement is willfully false in a material matter.

A plaintiff who supports that affidavit with a fake screenshot or a cloned-voice recording made with AI can be charged with perjury. A plaintiff or a defendant who testifies falsely at the hearing about fabricated evidence can be charged with perjury under G.L. c. 268, § 1. A defendant who also faces a criminal case should decide with counsel whether to testify at the restraining order hearing. The rules for authenticating digital evidence and challenging a screenshot or recording as fake are set out in Mass. G. Evid. § 901.

ChatGPT Conversations as Evidence in Court

ChatGPT conversations can be used against you in court. Police can obtain a person’s conversations with ChatGPT, Claude or another AI chatbot with a search warrant. Prosecutors can then offer the conversations as evidence at trial.

Massachusetts police obtain the same kind of records in two ways. The police can search a person’s phone or computer under a search warrant. The police can also serve a search warrant on the AI company under the federal Stored Communications Act, 18 U.S.C. § 2703. A search warrant must describe with particularity what the police may search for. A warrant that lets the police search an entire phone or account without limits can be challenged in a motion to suppress, as explained in the firm’s post on digital search warrants.

AI companies review some conversations themselves. OpenAI announced in August 2025 that human reviewers examine ChatGPT conversations flagged for plans to harm others. OpenAI stated that it may refer a conversation to law enforcement when its reviewers find an imminent threat of serious physical harm to others.

A person’s conversations with a chatbot about the person’s own case are not protected by the attorney-client privilege. In United States v. Heppner, No. 25-cr-503 (JSR) (S.D.N.Y. Feb. 17, 2026), Judge Jed S. Rakoff ruled that documents a criminal defendant created with the consumer version of Claude were not protected by the attorney-client privilege or the work product doctrine. Judge Rakoff ruled that Claude is not a lawyer, that the platform’s terms left the conversations without confidentiality, and that the defendant had not used Claude at his lawyers’ direction.

Massachusetts judges are not required to follow the Heppner decision. Massachusetts courts, however, apply the attorney-client privilege only to confidential communications between a client and a lawyer.

A person under investigation should put questions about the case to a defense lawyer, not to a chatbot. A person who deletes AI conversations after an investigation begins can face a separate charge. Under G.L. c. 268, § 13E, the Commonwealth can charge a person who alters, destroys or conceals a record with intent to impair the record’s use in a court proceeding.

AI Note Takers and Massachusetts Recording Law

Recording a meeting or phone call with an AI note taker, without the knowledge of every person in the conversation, can be charged as a felony in Massachusetts under the wiretap statute, G.L. c. 272, § 99. An employee who opens an AI transcription app during a meeting with a supervisor and tells no one in the room can face that charge.

Massachusetts is often called a two-party consent state. The statute prohibits secret recording, so the question in a criminal case is whether every participant knew about the recording. Section 99 defines an interception as secretly hearing or secretly recording the contents of a wire or oral communication with an intercepting device, by anyone other than a person given prior authority by all parties to the communication. The statute covers telephone calls and video calls as well as conversations in person.

If a defendant is convicted of an interception, the judge can impose up to five years in state prison or up to two and one half years in a house of correction, a fine of up to $10,000, or both.

In Commonwealth v. Hyde, 434 Mass. 594 (2001), the Supreme Judicial Court held that § 99 prohibits secret recording by private individuals. The Court also stated that a recording is not secret when the person recording announces the intention to record or holds the recording device in plain sight. An announcement at the start of a meeting or call that an AI note taker is recording gives every participant that knowledge.

The United States Court of Appeals for the First Circuit later held, in Project Veritas Action Fund v. Rollins, 982 F.3d 813 (1st Cir. 2020), that § 99 violates the First Amendment when it is applied to secret recording of police officers performing their duties in public. The Supreme Judicial Court noted that ruling in Commonwealth v. Rainey, 491 Mass. 632 (2023). The statute still applies to private meetings, workplace conversations and telephone calls. Home security cameras, workplace recordings and the civil claim under § 99 Q are covered on the firm’s page on the Massachusetts wiretap statute.

AI Fake ID Generators

Using an AI-generated image of a driver’s license can be charged as a felony under the Massachusetts fake ID statute, G.L. c. 90, § 24B. In 2024, the news site 404 Media reported that a website called OnlyFake sold AI-generated images of driver’s licenses and passports for $15 each. 404 Media reported that it used one of the images to pass an online identity check at a cryptocurrency exchange.

A person who uses an AI-made image of a driver’s license to open an account or pass an online age check can be charged under § 24B. Section 24B punishes a person who falsely makes a license to operate motor vehicles and a person who “in any way makes use of” a falsely made license. If a defendant is convicted under § 24B, the judge can impose up to five years in state prison, which makes the offense a felony.

A defendant charged over a digital image alone, with no physical card, can argue that an image on a screen is not a “license to operate motor vehicles” under § 24B. Many fake ID cases begin with a clerk-magistrate hearing rather than an arrest. The Registry of Motor Vehicles license suspension in a § 24B case is described in the firm’s fake ID FAQs for students.

Police Facial Recognition in Massachusetts

Massachusetts police can run a facial recognition search only by written request, and only under a court order or in an emergency. That rule is in G.L. c. 6, § 220, which took effect on July 1, 2021. Police officers use facial recognition software to identify a suspect from a photograph or a still image from video. Each law enforcement agency must document its facial recognition searches and report them to the Executive Office of Public Safety and Security every quarter.

When the police identified the defendant through a facial recognition search, defense counsel can move under Mass. R. Crim. P. 14 for the written request, the court order or the emergency justification, and the results of the search. Defense counsel can then challenge whether the police complied with § 220. Defense counsel can also challenge, through motions and cross-examination, whether any later identification of the defendant by a witness was reliable.

Defending a Criminal Charge Involving AI

Most AI cases depend on two questions, attribution and intent. The Commonwealth must prove that the defendant made or sent the image, message or recording, and that no one else used the defendant’s name, account or device to do it. The Commonwealth must also prove the mental state each statute requires, such as intent to harm under § 43A(b)(2) or intent to defraud under § 37E. When a case begins with an application for a criminal complaint, defense counsel can present the account and device evidence at the clerk-magistrate hearing, before the clerk-magistrate decides whether to issue a complaint.

A college or graduate student charged over AI conduct also faces a school disciplinary process that runs alongside the criminal case.

For a non-citizen, a conviction or a continuance without a finding on an AI-related charge can create immigration problems. An officer’s questions at an immigration interview about the arrest, or about conduct described in the police report, can create obstacles of their own. An immigration lawyer should review the exact question, the person’s status and the court record before the person files a petition or sits for an interview.

Attorney Joe Serpa has practiced Massachusetts criminal defense since 1995. He represents people charged with criminal harassment, identity fraud, restraining order violations and fake ID offenses in the district courts and the Boston Municipal Court. He also represents plaintiffs and defendants in 209A and 258E restraining order hearings.

Frequently Asked Questions

Related Serpa Law Office Resources

• AI deepfake and intimate image charges in Massachusetts
• Massachusetts criminal harassment defense
• Massachusetts revenge porn and nonconsensual intimate image law
• AI deepfake charges against Massachusetts students
• Violation of a 209A restraining order
• Coercive control and 209A restraining orders
• Digital search warrants for phones and computers
• Digital evidence in Massachusetts criminal cases
• The Massachusetts wiretap statute and secret recordings
• Massachusetts theft charges, including identity fraud
• Massachusetts fake ID defense
• Clerk-magistrate hearings in Massachusetts
• Immigration consequences of Massachusetts criminal charges