AI-Generated Sexual Images Are Creating New Criminal-Evidence Problems in Georgia
A sexually explicit image can look convincing enough to trigger an accusation before anyone knows how it was created. Generative artificial intelligence now allows someone to place a real person’s face onto a fabricated body, alter an existing photograph, or create a realistic sexual image depicting an event that never happened. Once an image like that reaches police, social media, a school, or an employer, the first question can quickly become who made it and why.
Georgia law already reaches certain falsely created sexual images, while lawmakers continue considering additional ways to address generative AI. For someone accused of creating, possessing, transmitting, or distributing synthetic sexual material, the image itself can be only the beginning of the evidence. Guidance from an experienced Atlanta sex crime lawyer can help examine how the file was created, where it came from, who controlled the relevant accounts and devices, and whether prosecutors can reliably connect the image to the person they have accused.
Georgia Law Already Covers Some Falsely Created Sexual Images
Artificially created intimate images are not entirely outside Georgia’s existing criminal law. O.C.G.A. § 16-11-90 addresses the electronic transmission or posting of nude or sexually explicit images of an adult without the depicted person’s consent under specified circumstances. The statute expressly includes a “falsely created videographic or still image.”
That language was added before today’s generative-AI tools became widely available. It means a prosecution does not necessarily depend on proving that the depicted person actually posed nude or participated in the sexual conduct shown. A manipulated or fabricated image can fall within the statute when the other required elements are satisfied.
The statute also contains additional requirements concerning the transmission, the absence of consent, harassment or financial loss, and whether the transmission serves a legitimate purpose to the depicted person. The existence of an AI-generated sexual image alone does not resolve every element of a criminal charge. Investigators still have to determine who created or transmitted the material and what occurred when it was shared.
Determining Whether an Image Is Real Can Become a Major Issue
A realistic image no longer proves that the event depicted actually happened. Generative AI has weakened one of the assumptions people commonly make about photographs and videos: that seeing something provides reliable evidence that it occurred.
A disputed file can require closer examination of its origin. The original image may contain information about when it was created or modified. Devices can contain earlier versions, source photographs, downloaded files, editing applications, prompts, cached material, or communications discussing the image. A screenshot or reposted copy may strip away information that existed in the original file.
Visual imperfections can sometimes raise questions about manipulation, but appearance alone is a poor substitute for examining the underlying evidence. AI systems continue improving, while compression, filters, ordinary photo editing, and repeated social-media uploads can also alter legitimate photographs. A defense based solely on claiming that an image “looks fake” leaves too much unanswered.
Authentication therefore takes on added importance. Under O.C.G.A. § 24-9-901, evidence must be supported sufficiently for a finding that it is what the party offering it claims. In an AI-image case, the dispute can involve both what the file depicts and what the prosecution claims the file represents.
Finding the Image Does Not Establish Who Created It
Source identification can become one of the hardest parts of a synthetic-image investigation. An image found on someone’s phone may have been generated there, received in a message, downloaded from another account, automatically saved, or forwarded through a group conversation.
The same problem arises with online accounts. A username or profile associated with a person does not by itself explain who had access to the account when a particular file was created or sent. Investigators may look at login records, linked email addresses, phone numbers, device information, passwords, IP records, and communications surrounding the image.
Creation and distribution can also involve different people. One person might generate a synthetic image and send it privately to another person, who later posts it publicly. Someone else may modify the original before forwarding it again. Once multiple copies begin circulating, investigators have to reconstruct the path instead of assuming that the person possessing one copy started the chain.
Messages discussing the image can be particularly significant. A conversation about making a deepfake, requesting a particular edit, reacting to a finished image, or planning to distribute it can provide evidence about authorship or intent. The absence of such communications can leave prosecutors relying more heavily on device and account evidence.
Synthetic Images Involving Minors Raise Additional Concerns
The legal stakes become especially serious when an image appears to depict a child. Georgia separately regulates sexually explicit depictions of minors through its child-exploitation laws, and lawmakers have repeatedly considered how those laws should address computer-generated material.
The question of what the image actually depicts can become complicated when AI is involved. A synthetic image might use the face of an identifiable minor, combine photographs of several people, transform an adult to appear younger, or depict an entirely computer-generated person. Those distinctions can affect which law prosecutors attempt to apply and what they must prove.
Georgia lawmakers considered HB 171 during the 2025-2026 legislative session. Among other provisions, the proposal sought to prohibit distribution of computer-generated obscene material depicting a child and addressed enhanced sentencing when artificial intelligence is used in certain crimes. The House passed a version of the measure in 2025, and the legislation remained under consideration in the Senate during 2026.
The legislative activity reflects a broader problem for criminal courts. Existing statutes were often written around photographs or videos created through conventional means, while generative AI can produce material without a camera, a physical encounter, or even a real person posing for the underlying image.
Georgia Lawmakers Considered a New “Virtual Peeping” Offense in 2026
Georgia legislators also addressed synthetic images more directly through Senate Bill 398 during the 2026 session. The bill proposed criminal offenses for “virtual peeping” and sought to prohibit specified uses of generative artificial intelligence to create images of individuals when the person generating the image knew that the creation lacked authorization or consent.
SB 398 received substantial support in both chambers. The Senate passed a substitute version 48-1 on February 20, 2026, and the House later passed its own substitute 164-2 on March 31. The Senate disagreed with the House substitute that same day, so the proposal did not emerge from the regular session in an agreed final form.
Its progress still illustrates how quickly Georgia’s criminal laws are being forced to respond to generative AI. The existing prohibition on certain falsely created sexual images focuses substantially on transmission or posting. The proposed virtual-peeping legislation addressed the use of AI to generate unauthorized images more directly.
For defendants, rapid legislative change also makes the timing of alleged conduct important. A prosecutor cannot simply apply a later criminal prohibition to earlier conduct. The statute in effect when the alleged act occurred, along with the exact conduct prosecutors claim took place, has to guide the analysis.
Original Files Can Matter More Than Social-Media Copies
AI-generated sexual images often reach investigators after being reposted, screenshotted, compressed, or forwarded numerous times. By then, important information can be missing.
The original file can provide a much better starting point. Investigators may examine metadata where available, creation dates, modification history, associated application data, or copies stored elsewhere on the device. If the allegation involves an AI image generator, account history or communications can sometimes help establish how the image was produced.
A social-media screenshot creates different limitations. It may identify who posted the file without establishing who generated it. It can also conceal the filename, original dimensions, earlier versions, or information associated with the source file.
Device history can help distinguish creation from receipt. A phone containing only an incoming message with the image presents a different evidentiary record from a device containing source photographs, editing activity, saved drafts, and communications directing changes to the image. Those differences can determine whether investigators actually have evidence of creation or merely evidence of possession.
AI Evidence Makes Preservation and Authentication More Important
The ability to manufacture realistic content increases the importance of preserving the evidence in its original form. If an accuser provides only a screenshot while deleting the underlying message, account history, or original file, the defense may lose information that could have helped determine where the image came from.
The same concern applies to devices seized by police. A forensic extraction can preserve app records, account information, images, messages, and other artifacts before ordinary phone activity changes what remains. Records held by online platforms may also have limited retention periods.
In a case involving synthetic sexual material, the source file and the digital trail surrounding it can carry more evidentiary weight than the disturbing image investigators see first. Working with an Atlanta sex crime lawyer can help identify the difference between evidence showing that an image existed and evidence establishing who generated, altered, possessed, or distributed it.
Contact The Spizman Firm Today
AI-generated sexual images can create accusations that move faster than the underlying evidence. A realistic deepfake can provoke an immediate reaction, yet a criminal case still requires proof connecting the material to particular conduct and a particular person. Devices, account records, original files, messages, and the applicable version of Georgia law all deserve careful examination.
If you are being investigated or have been charged in Atlanta or elsewhere in Georgia in a case involving deepfakes, synthetic sexual images, or other digital evidence, contact The Spizman Firm today to speak with an experienced Atlanta sex crime lawyer and learn how we can examine the source and authenticity of the disputed material and challenge unsupported conclusions about who created or distributed it.
Sources:
- Georgia Code § 16-11-90 — Prohibition on Nude or Sexually Explicit Electronic Transmissions
law.justia.com/codes/georgia/title-16/chapter-11/article-3/part-3/section-16-11-90/ - Georgia Code § 24-9-901 — Requirement of Authentication or Identification
law.justia.com/codes/georgia/title-24/chapter-9/article-1/section-24-9-901/ - Georgia General Assembly — SB 398, Virtual Peeping and Generative Artificial Intelligence
legis.ga.gov/legislation/72204 - Georgia General Assembly — HB 171, Computer-Generated Obscene Material Depicting a Child and AI Sentencing Provisions
legis.ga.gov/legislation/69648 - Georgia General Assembly — SB 9, Ensuring Accountability for Illegal AI Activities Act
legis.ga.gov/api/legislation/document/20252026/229878