Seventh Circuit Refuses to Stretch Existing Law to Reach AI-Generated Content: United States v. Anderegg
September 3, 2026
By: Alex C. Wimmer
Artificial intelligence continues to test legal frameworks written before modern generative AI existed. Companies utilizing AI need to pay close attention to how Court’s are interpreting and applying laws written before AI became mainstream in order to understand how it may affect them. In United States v. Anderegg, No. 25-1354, the Seventh Circuit confronted that problem directly, recognizing that generative AI can now create images of “virtual children” that are “virtually indistinguishable” from images depicting actual abuse, while also emphasizing that lower courts remain bound by existing Supreme Court precedent. United States v. Anderegg, No. 25-1354, slip op. at 3 (7th Cir. Aug. 25, 2026).
The decision is important not because the court minimized concerns about AI-generated child sexual abuse material (“CSAM”), but because it appears to reflect a reluctance to resolve those concerns by expanding preexisting law beyond the limits imposed by Supreme Court doctrine. The court expressly acknowledged “some concerns about the lines” drawn by Stanley v. Georgia, Osborne v. Ohio, and Ashcroft v. Free Speech Coalition, but emphasized that it was “not free to redraw them” itself. Anderegg, slip op. at 3.
Background
According to the Seventh Circuit, Meta Platforms reported the online transmission of potential CSAM to the CyberTipline of the National Center for Missing and Exploited Children. The opinion states that an Instagram user apparently sent a direct message containing what appeared to be AI-generated images depicting CSAM to an account belonging to a minor, and that the corresponding message history showed the user explaining how he created the images by entering text prompts into Stable Diffusion, a text-to-image generative AI model. Anderegg, slip op. at 3-4.
The government alleged that Anderegg installed Stable Diffusion on his personal laptop, added software components that allowed the program to create depictions of human genitalia, and used curated text prompts and negative prompts to generate photo-realistic images of minors in sexually explicit positions. The government also conceded that the images charged did not depict, and could not be linked to, an actual child. Anderegg, slip op. at 4 & n.2.
The indictment charged Anderegg with producing, distributing, and transferring obscene visual depictions of minors under 18 U.S.C. § 1466A(a)(1), (d)(1), and with possessing at least one obscene visual depiction of a minor under 18 U.S.C. § 1466A(b)(1), (d)(4). The district court dismissed only the possession count, leaving the production, distribution, and transfer counts intact. Anderegg, slip op. at 5. The government then filed an interlocutory appeal challenging only the dismissal of the possession charge. Id. at 3.
The Seventh Circuit's Decision
The Seventh Circuit affirmed. Reviewing the as-applied constitutional ruling de novo, the court framed the question narrowly: “does the First Amendment allow the government to proscribe the in-home possession of obscene virtual CSAM.” Anderegg, slip op. at 17. The court answered no. It concluded that because Anderegg was accused of possessing obscene materials at home, his circumstances fell within Stanley; and because the government’s arguments for removing the materials from Stanley’s protection had been “squarely rejected” by the Supreme Court in Free Speech Coalition, § 1466A(b)(1) was unconstitutional as applied. Anderegg, slip op. at 23-24.
The court did not hold that all AI-related conduct involving such images is constitutionally protected. To the contrary, its opinion emphasized the narrow posture of the appeal and noted that the remaining charges, including counts alleging production, distribution, and transfer to a minor, were not before it. Anderegg, slip op. at 5, 16-17.
Why the Decision Matters
The significance of Anderegg extends beyond the specific context of AI-generated sexual images.
The facts presented a difficult case for a restrained approach. The panel described the government’s disturbing allegations involving Stable Diffusion, added software components, and curated prompts used to generate photo-realistic images of minors in sexually explicit positions. Anderegg, slip op. at 4. The court also acknowledged that current generative AI technology complicates the distinction between virtual and actual child sexual abuse material. Id. at 3.
Nevertheless, the Seventh Circuit declined to stretch existing doctrine to address new technological risks. The court rejected the government’s attempt to distinguish Stanley based on § 1466A’s interstate-commerce element, explaining that the jurisdictional requirement did not change the fact that the government was using the statute to punish in-home possession. Anderegg, slip op. at 15-18. It also rejected the government’s argument that Stanley should be limited to obscene materials involving adults, reasoning that Stanley turned on location, “the privacy of his own home,” not on the content category the government proposed. Anderegg, slip op. at 18-19.
The most important part of the decision may be the court’s treatment of the government’s AI-specific policy arguments. The government argued that virtual AI-generated CSAM could be used to groom minors, normalize sexual abuse, reduce law enforcement’s ability to distinguish real from virtual images, and support the market for actual CSAM. The court did not dismiss those concerns. But it held that the same or substantially similar arguments had been addressed and rejected by the Supreme Court in Free Speech Coalition, and that the Seventh Circuit was “dutybound” to follow that precedent unless and until the Supreme Court revisits it. Anderegg, slip op. at 20-24.
A Broader Lesson for AI Litigation
Anderegg may offer a broader lesson for AI litigation. Courts may be sympathetic to arguments that AI creates new risks that older statutes and precedents did not anticipate. But sympathy for those concerns does not necessarily permit a court to expand a statute or constitutional exception beyond its existing bounds.
The court’s concurrence underscores that point. Judge Lee, joined by Judge Kolar, wrote separately to state that, given “significant advances in computer-generated image technology, particularly in the field of artificial intelligence,” courts would benefit from further Supreme Court guidance on the intersection between the First Amendment and virtual CSAM. Anderegg, slip op. at 25 (Lee, J., concurring). The concurrence recognized that AI-generated CSAM may now be “virtually indistinguishable” from images created through the abuse of an actual child and may impose increasing burdens on law enforcement, but concluded that the record was insufficient and that the panel remained bound by Free Speech Coalition. Anderegg, slip op. at 25-27 (Lee, J., concurring).
That reasoning may matter well beyond the criminal statutes at issue in Anderegg. In disputes involving AI outputs, plaintiffs, regulators, and prosecutors often argue that existing legal regimes should be read broadly because AI creates harms that lawmakers did not foresee. Anderegg suggests that courts may insist on a clear statutory hook and a valid doctrinal basis before extending pre-AI rules to AI-generated conduct.
In practical terms, the decision supports a disciplined approach to AI regulation and AI litigation: identify the conduct, identify the statute or doctrine that allegedly reaches it, and then ask whether that law actually applies as written. Anderegg indicates that a court may not be willing to fill perceived regulatory gaps simply because the underlying technology is troubling or the alleged conduct is difficult to defend.
Potential Implications Beyond Criminal Law
Although Anderegg arises in a criminal context, its reasoning could influence disputes involving:
- AI-generated content and intellectual property rights;
- consumer protection claims involving AI outputs;
- AI-driven employment decisions;
- deepfake and synthetic media regulation;
- privacy and data-use disputes; and
- emerging AI governance frameworks.
In each of those areas, parties may argue that existing laws should be interpreted broadly to reach AI-related conduct. Courts may increasingly confront the same tension presented in Anderegg: whether perceived gaps in the law should be filled by judicial interpretation or by legislative action.
Conclusion
United States v. Anderegg is noteworthy because it appears to show judicial restraint in the face of troubling AI-enabled conduct. The Seventh Circuit did not say the concerns were imaginary. It said that the governing precedent controlled, that the record did not permit the court to depart from that precedent, and that lower courts cannot redraw constitutional lines simply because technology has changed. Anderegg, slip op. at 3, 23-27.
For companies developing, deploying, or regulating AI systems, the decision is a reminder that legal risk analysis should begin with the text of the governing statute and the controlling doctrine. Policy concerns about AI may drive legislative change, enforcement priorities, or future Supreme Court review, but Anderegg suggests that at least some courts will resist converting those concerns into liability where the existing law does not reach the conduct at issue.
Disclaimer: The contents of this article should not be construed as legal advice or a legal opinion on any specific facts or circumstances. The contents are intended for general informational purposes only, and you are urged to consult with counsel concerning your situation and specific legal questions you may have.
Practices
September 3, 2026
By: Alex C. Wimmer
Artificial intelligence continues to test legal frameworks written before modern generative AI existed. Companies utilizing AI need to pay close attention to how Court’s are interpreting and applying laws written before AI became mainstream in order to understand how it may affect them. In United States v. Anderegg, No. 25-1354, the Seventh Circuit confronted that problem directly, recognizing that generative AI can now create images of “virtual children” that are “virtually indistinguishable” from images depicting actual abuse, while also emphasizing that lower courts remain bound by existing Supreme Court precedent. United States v. Anderegg, No. 25-1354, slip op. at 3 (7th Cir. Aug. 25, 2026).
The decision is important not because the court minimized concerns about AI-generated child sexual abuse material (“CSAM”), but because it appears to reflect a reluctance to resolve those concerns by expanding preexisting law beyond the limits imposed by Supreme Court doctrine. The court expressly acknowledged “some concerns about the lines” drawn by Stanley v. Georgia, Osborne v. Ohio, and Ashcroft v. Free Speech Coalition, but emphasized that it was “not free to redraw them” itself. Anderegg, slip op. at 3.
Background
According to the Seventh Circuit, Meta Platforms reported the online transmission of potential CSAM to the CyberTipline of the National Center for Missing and Exploited Children. The opinion states that an Instagram user apparently sent a direct message containing what appeared to be AI-generated images depicting CSAM to an account belonging to a minor, and that the corresponding message history showed the user explaining how he created the images by entering text prompts into Stable Diffusion, a text-to-image generative AI model. Anderegg, slip op. at 3-4.
The government alleged that Anderegg installed Stable Diffusion on his personal laptop, added software components that allowed the program to create depictions of human genitalia, and used curated text prompts and negative prompts to generate photo-realistic images of minors in sexually explicit positions. The government also conceded that the images charged did not depict, and could not be linked to, an actual child. Anderegg, slip op. at 4 & n.2.
The indictment charged Anderegg with producing, distributing, and transferring obscene visual depictions of minors under 18 U.S.C. § 1466A(a)(1), (d)(1), and with possessing at least one obscene visual depiction of a minor under 18 U.S.C. § 1466A(b)(1), (d)(4). The district court dismissed only the possession count, leaving the production, distribution, and transfer counts intact. Anderegg, slip op. at 5. The government then filed an interlocutory appeal challenging only the dismissal of the possession charge. Id. at 3.
The Seventh Circuit's Decision
The Seventh Circuit affirmed. Reviewing the as-applied constitutional ruling de novo, the court framed the question narrowly: “does the First Amendment allow the government to proscribe the in-home possession of obscene virtual CSAM.” Anderegg, slip op. at 17. The court answered no. It concluded that because Anderegg was accused of possessing obscene materials at home, his circumstances fell within Stanley; and because the government’s arguments for removing the materials from Stanley’s protection had been “squarely rejected” by the Supreme Court in Free Speech Coalition, § 1466A(b)(1) was unconstitutional as applied. Anderegg, slip op. at 23-24.
The court did not hold that all AI-related conduct involving such images is constitutionally protected. To the contrary, its opinion emphasized the narrow posture of the appeal and noted that the remaining charges, including counts alleging production, distribution, and transfer to a minor, were not before it. Anderegg, slip op. at 5, 16-17.
Why the Decision Matters
The significance of Anderegg extends beyond the specific context of AI-generated sexual images.
The facts presented a difficult case for a restrained approach. The panel described the government’s disturbing allegations involving Stable Diffusion, added software components, and curated prompts used to generate photo-realistic images of minors in sexually explicit positions. Anderegg, slip op. at 4. The court also acknowledged that current generative AI technology complicates the distinction between virtual and actual child sexual abuse material. Id. at 3.
Nevertheless, the Seventh Circuit declined to stretch existing doctrine to address new technological risks. The court rejected the government’s attempt to distinguish Stanley based on § 1466A’s interstate-commerce element, explaining that the jurisdictional requirement did not change the fact that the government was using the statute to punish in-home possession. Anderegg, slip op. at 15-18. It also rejected the government’s argument that Stanley should be limited to obscene materials involving adults, reasoning that Stanley turned on location, “the privacy of his own home,” not on the content category the government proposed. Anderegg, slip op. at 18-19.
The most important part of the decision may be the court’s treatment of the government’s AI-specific policy arguments. The government argued that virtual AI-generated CSAM could be used to groom minors, normalize sexual abuse, reduce law enforcement’s ability to distinguish real from virtual images, and support the market for actual CSAM. The court did not dismiss those concerns. But it held that the same or substantially similar arguments had been addressed and rejected by the Supreme Court in Free Speech Coalition, and that the Seventh Circuit was “dutybound” to follow that precedent unless and until the Supreme Court revisits it. Anderegg, slip op. at 20-24.
A Broader Lesson for AI Litigation
Anderegg may offer a broader lesson for AI litigation. Courts may be sympathetic to arguments that AI creates new risks that older statutes and precedents did not anticipate. But sympathy for those concerns does not necessarily permit a court to expand a statute or constitutional exception beyond its existing bounds.
The court’s concurrence underscores that point. Judge Lee, joined by Judge Kolar, wrote separately to state that, given “significant advances in computer-generated image technology, particularly in the field of artificial intelligence,” courts would benefit from further Supreme Court guidance on the intersection between the First Amendment and virtual CSAM. Anderegg, slip op. at 25 (Lee, J., concurring). The concurrence recognized that AI-generated CSAM may now be “virtually indistinguishable” from images created through the abuse of an actual child and may impose increasing burdens on law enforcement, but concluded that the record was insufficient and that the panel remained bound by Free Speech Coalition. Anderegg, slip op. at 25-27 (Lee, J., concurring).
That reasoning may matter well beyond the criminal statutes at issue in Anderegg. In disputes involving AI outputs, plaintiffs, regulators, and prosecutors often argue that existing legal regimes should be read broadly because AI creates harms that lawmakers did not foresee. Anderegg suggests that courts may insist on a clear statutory hook and a valid doctrinal basis before extending pre-AI rules to AI-generated conduct.
In practical terms, the decision supports a disciplined approach to AI regulation and AI litigation: identify the conduct, identify the statute or doctrine that allegedly reaches it, and then ask whether that law actually applies as written. Anderegg indicates that a court may not be willing to fill perceived regulatory gaps simply because the underlying technology is troubling or the alleged conduct is difficult to defend.
Potential Implications Beyond Criminal Law
Although Anderegg arises in a criminal context, its reasoning could influence disputes involving:
- AI-generated content and intellectual property rights;
- consumer protection claims involving AI outputs;
- AI-driven employment decisions;
- deepfake and synthetic media regulation;
- privacy and data-use disputes; and
- emerging AI governance frameworks.
In each of those areas, parties may argue that existing laws should be interpreted broadly to reach AI-related conduct. Courts may increasingly confront the same tension presented in Anderegg: whether perceived gaps in the law should be filled by judicial interpretation or by legislative action.
Conclusion
United States v. Anderegg is noteworthy because it appears to show judicial restraint in the face of troubling AI-enabled conduct. The Seventh Circuit did not say the concerns were imaginary. It said that the governing precedent controlled, that the record did not permit the court to depart from that precedent, and that lower courts cannot redraw constitutional lines simply because technology has changed. Anderegg, slip op. at 3, 23-27.
For companies developing, deploying, or regulating AI systems, the decision is a reminder that legal risk analysis should begin with the text of the governing statute and the controlling doctrine. Policy concerns about AI may drive legislative change, enforcement priorities, or future Supreme Court review, but Anderegg suggests that at least some courts will resist converting those concerns into liability where the existing law does not reach the conduct at issue.
Disclaimer: The contents of this article should not be construed as legal advice or a legal opinion on any specific facts or circumstances. The contents are intended for general informational purposes only, and you are urged to consult with counsel concerning your situation and specific legal questions you may have.