AI Regulation Tracker / State legislation
Maryland Makes Knowing Use of AI Deepfakes for Identity Fraud a Crime
Maryland Senate Bill 8, "Identity Fraud - Artificial Intelligence and Deepfake Representations," was approved by the Governor on May 12, 2026 as Chapter 445 and takes effect October 1, 2026. It creates criminal identity-fraud liability for knowingly using AI or deepfakes with fraudulent intent to impersonate or harm another person, with penalties reaching 10 years and $15,000 when there are multiple victims. This is a live statute with a fixed activation date, not a proposal.
Maryland has moved AI and deepfakes squarely into its criminal identity-fraud law. Senate Bill 8 of the 2026 Regular Session, titled "Identity Fraud - Artificial Intelligence and Deepfake Representations," was approved by the Governor on May 12, 2026 and carried into the code as Chapter 445. On the General Assembly's own bill page the status line reads "Approved by the Governor - Chapter 445." The law takes effect on October 1, 2026.
What the statute actually does
The core move is simple and deliberate. Maryland already had an identity-fraud statute that punished using another person's identifying information to defraud or to represent yourself as someone you are not. SB 8 pulls AI-generated content and deepfakes into that same framework. It makes it a crime to knowingly use artificial intelligence or a deepfake representation, with fraudulent intent, to impersonate another person or to cause harm.
Two words are doing the heavy lifting: "knowingly" and "fraudulent intent." This is not a strict-liability rule that sweeps in anyone who ever made a synthetic video. It targets the person who knows what they are doing and does it to deceive, to impersonate, or to injure. That intent requirement is what keeps the statute aimed at bad actors rather than at ordinary creative or business uses of the technology.
The penalty structure scales with how many people got hurt. Where the offense involves a single victim, it carries a term of up to 5 years in prison and a fine of up to $10,000. Where there are multiple victims, the ceiling rises to up to 10 years and a fine of up to $15,000. Those are real criminal exposures, not civil penalties, and they attach to the conduct of using the AI or the deepfake, not merely to building it.
Who this reaches
The practical reach is broad because the technology is broad. Anyone deploying deepfakes, face-swaps, synthetic video, or voice-cloning in Maryland is now operating against a criminal backstop. That includes the obvious cases, the scammer who clones an executive's voice to authorize a wire transfer, the fraudster who spins up a fake video of a real person to trick a victim, and it includes less obvious ones where a synthetic likeness is used to impersonate someone in a transaction or a communication meant to deceive.
It is worth being clear about what the statute does not do. It does not ban AI, it does not ban deepfakes as a category, and it does not create liability for lawful, disclosed, or consented uses. The line the legislature drew is intent-based. Use the tools to defraud or impersonate, and you are exposed. Use them openly and honestly, and the statute is not aimed at you.
Do not confuse this with Maryland's insurer AI bill
There is a second Maryland AI measure moving in the same period, and it is easy to run the two together. House Bill 1563 deals with insurers and their use of AI, including reporting expectations around how AI systems are used in the insurance context. That is a different subject, a different chamber of origin, and a different mechanism. SB 8, the subject of this briefing, is a criminal identity-fraud statute about deepfakes and impersonation. HB 1563 is an insurance-regulatory measure about AI reporting. If you are tracking Maryland AI law, keep them in separate columns. They do not amend each other and they answer different questions.
Where this sits in the national picture
SB 8 is another data point in a wave of state deepfake-liability laws. States have been building criminal and civil exposure around synthetic media for a couple of years now, hitting election deepfakes, non-consensual intimate imagery, and fraud one bill at a time, because there is no single federal statute that does the whole job. Maryland's contribution is a net-new state criminal statute that plugs AI and deepfakes into an existing identity-fraud framework rather than standing up a brand-new offense from scratch. That is a common and durable drafting pattern, and it makes the law easier for prosecutors to use because they already know how to charge identity fraud.
For anyone operating across state lines, the takeaway is that the deepfake-liability map is filling in, and the rules are not identical from state to state. What is a disclosure question in one state can be a criminal-intent question in another. Maryland has now planted a clear intent-based criminal marker with a fixed October 1, 2026 start.
What to do before October 1, 2026
If you build or deploy synthetic media that could touch Maryland, treat the effective date as a real deadline. Confirm that any use of AI likenesses, voice-cloning, or deepfake representations in your products or campaigns is consented, disclosed where appropriate, and never used to impersonate a real person in a way meant to deceive. Tighten the controls that keep your tools from being used for fraudulent impersonation, since the statute reaches the knowing, intentional use of these representations. Brief your legal and compliance teams on the single-victim and multiple-victim penalty tiers so the stakes are understood internally. And keep SB 8 separate from HB 1563 in your tracking so you do not misread which obligations apply to whom. None of this is legal advice, and how the statute applies to a specific product or situation is a question for Maryland counsel, but the direction is unambiguous: knowing, fraudulent deepfake impersonation is now a crime in Maryland as of October 1, 2026.
Questions professionals are asking
Is Maryland SB 8 actually law, or just a proposal?
It is law. Senate Bill 8 was approved by the Governor on May 12, 2026 and enrolled as Chapter 445. The General Assembly bill page shows the status "Approved by the Governor - Chapter 445." It becomes effective on October 1, 2026.
What conduct does it criminalize?
Knowingly using artificial intelligence or a deepfake representation, with fraudulent intent, to impersonate another person or to cause harm. It folds AI and deepfakes into Maryland's existing identity-fraud framework. The intent requirement means it targets deliberate deception, not ordinary or consented uses of the technology.
What are the penalties?
They scale with the number of victims. A single-victim offense carries up to 5 years in prison and a fine of up to $10,000. A multiple-victim offense carries up to 10 years and a fine of up to $15,000. These are criminal penalties.
How is this different from Maryland HB 1563?
HB 1563 addresses insurers and AI reporting in the insurance context. SB 8 is a criminal identity-fraud statute about deepfakes and impersonation. They are separate measures on separate subjects and do not amend each other.
Who needs to pay attention before October 1, 2026?
Anyone deploying deepfakes, synthetic media, or voice-cloning that could touch Maryland, along with businesses and platforms whose tools or users reach Maryland residents. Confirm consent, disclosure, and controls that prevent your technology from being used for fraudulent impersonation, and consult Maryland counsel on your specific situation.
RELATED BRIEFINGS
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any statute or requirement applies to your situation with qualified counsel in the relevant jurisdiction.