WHAT THIS BILL REGULATES · 1 REQUIREMENT TYPE
How Is This Bill Enforced
Verbatim statutory text on the left; plain-language analysis and a per-section checklist on the right. Numbered markers cross-link to the matching checklist row.
(a)(1)–(6) DEFINITIONS.—In this section: (1) ARTIFICIAL INTELLIGENCEartificial intelligenceThe term 'artificial intelligence' has the meaning given the term in section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401).17 U.S.C. § 514(a)(1).—The term 'artificial intelligenceartificial intelligenceThe term 'artificial intelligence' has the meaning given the term in section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401).17 U.S.C. § 514(a)(1)' has the meaning given the term in section 5002 of the National Artificial Intelligenceartificial intelligenceThe term 'artificial intelligence' has the meaning given the term in section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401).17 U.S.C. § 514(a)(1) Initiative Act of 2020 (15 U.S.C. 9401). (2) ARTIFICIAL INTELLIGENCE MODELartificial intelligence modelThe term 'artificial intelligence model' means a component of an information system that implements artificial intelligence technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs.17 U.S.C. § 514(a)(2).—The term 'artificial intelligence modelartificial intelligence modelThe term 'artificial intelligence model' means a component of an information system that implements artificial intelligence technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs.17 U.S.C. § 514(a)(2)' means a component of an information system that implements artificial intelligenceartificial intelligenceThe term 'artificial intelligence' has the meaning given the term in section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401).17 U.S.C. § 514(a)(1) technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs. (3) DEVELOPERdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3).—The term 'developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3)'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4) for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence modelartificial intelligence modelThe term 'artificial intelligence model' means a component of an information system that implements artificial intelligence technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs.17 U.S.C. § 514(a)(2); or (II) the use of the training dataset to train the artificial intelligence modelartificial intelligence modelThe term 'artificial intelligence model' means a component of an information system that implements artificial intelligence technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs.17 U.S.C. § 514(a)(2); and (B) does not include a noncommercial end user of a generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4). (4) GENERATIVE ARTIFICIAL INTELLIGENCE MODELgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4).—The term 'generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4)'— (A) means an artificial intelligence modelartificial intelligence modelThe term 'artificial intelligence model' means a component of an information system that implements artificial intelligence technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs.17 U.S.C. § 514(a)(2) that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence modelartificial intelligence modelThe term 'artificial intelligence model' means a component of an information system that implements artificial intelligence technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs.17 U.S.C. § 514(a)(2) described in subparagraph (A), even if created by a third party. (5) SUBSTANTIALLY MODIFYsubstantially modifyThe term 'substantially modify', with respect to a generative artificial intelligence model, means to take 1 or more actions leading to a new version of, new release of, or other update to the generative artificial intelligence model that materially changes the functionality or performance of the generative artificial intelligence model, including by retraining or fine tuning the generative artificial intelligence model.17 U.S.C. § 514(a)(5).—The term 'substantially modifysubstantially modifyThe term 'substantially modify', with respect to a generative artificial intelligence model, means to take 1 or more actions leading to a new version of, new release of, or other update to the generative artificial intelligence model that materially changes the functionality or performance of the generative artificial intelligence model, including by retraining or fine tuning the generative artificial intelligence model.17 U.S.C. § 514(a)(5)', with respect to a generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4), means to take 1 or more actions leading to a new version of, new release of, or other update to the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4) that materially changes the functionality or performance of the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4), including by retraining or fine tuning the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4). (6) TRAINING MATERIALtraining materialThe term 'training material' means individual works or components thereof used for the purpose of training a generative artificial intelligence model, including a combination of text, images, audio, or other categories of expressive materials, as well as annotations describing the material.17 U.S.C. § 514(a)(6).—The term 'training materialtraining materialThe term 'training material' means individual works or components thereof used for the purpose of training a generative artificial intelligence model, including a combination of text, images, audio, or other categories of expressive materials, as well as annotations describing the material.17 U.S.C. § 514(a)(6)' means individual works or components thereof used for the purpose of training a generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4), including a combination of text, images, audio, or other categories of expressive materials, as well as annotations describing the material.
This subsection establishes the defined terms used throughout new Section 514. Key definitions include developer — which requires both creation or substantial modification of a generative AI model and engagement in or supervision of training dataset curation or use — and generative artificial intelligence model, which extends to subsequent variations even if created by third parties. Noncommercial end users are expressly excluded from the developer definition.
(b)(1) IN GENERAL.—The legal or beneficial owner of an exclusive right under a copyright, or a person authorized to act on the owner's behalf, may request the clerk of any United States district court to issue a subpoena to a developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) for disclosure of copies of, or records sufficient to identify with certainty, the copyrighted works, or any portion thereof, likely owned or controlled by the legal or beneficial owner that were used by the developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) to train the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4), if the legal or beneficial owner or authorized person has a subjective good faith belief that the developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) used some or all of 1 or more such copyrighted works to train the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4).
(b)(2) SUBPOENA LIMITED TO REQUESTER'S OWN COPYRIGHTED WORK.—Nothing in paragraph (1) shall be construed to authorize a legal or beneficial owner of an exclusive right under a copyright, or a person authorized to act on the owner's behalf, to request a subpoena for disclosure of copies of, or records sufficient to identify with certainty, copyrighted works likely owned or controlled by any person other than the legal or beneficial owner.
This subsection establishes the core mechanism of the TRAIN Act: a copyright owner (or authorized representative) may request the clerk of any U.S. district court to issue a subpoena compelling a developer to disclose copies of, or records sufficient to identify, the copyrighted works used to train its generative AI model. The request requires a subjective good-faith belief that some or all of the requester's copyrighted works were used in training. Paragraph (2) expressly limits the subpoena to the requester's own works — it cannot be used to discover third-party copyrighted works in a developer's training data.
(c) CONTENTS OF REQUEST.—A request under subsection (b) may be made by filing with the clerk— (1) a proposed subpoena; and (2) a sworn declaration to the effect that— (A) the legal or beneficial owner or authorized person has a subjective good faith belief that the developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) used some or all of 1 or more of the copyrighted works owned or controlled by the legal or beneficial owner to train the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4); (B) the purpose for which the subpoena is sought is to obtain copies of the training materialtraining materialThe term 'training material' means individual works or components thereof used for the purpose of training a generative artificial intelligence model, including a combination of text, images, audio, or other categories of expressive materials, as well as annotations describing the material.17 U.S.C. § 514(a)(6), or records sufficient to identify with certainty the training materialtraining materialThe term 'training material' means individual works or components thereof used for the purpose of training a generative artificial intelligence model, including a combination of text, images, audio, or other categories of expressive materials, as well as annotations describing the material.17 U.S.C. § 514(a)(6), used to train the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4) in order to determine whether the developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) has used copyrighted works owned or controlled by the legal or beneficial owner in connection with the generative artificial intelligence modelgenerative artificial intelligence modelThe term 'generative artificial intelligence model'— (A) means an artificial intelligence model that emulates the structure and characteristics of input data in order to generate derived synthetic content, which may include images, videos, audio, text, and other digital content; and (B) includes any subsequent variation on an artificial intelligence model described in subparagraph (A), even if created by a third party.17 U.S.C. § 514(a)(4); and (C) the copies or records will only be used for the purpose of protecting the rights of the legal or beneficial owner.
This subsection specifies the formal requirements for filing a subpoena request. The requester must submit a proposed subpoena and a sworn declaration attesting to three things: (1) a subjective good-faith belief that the developer used the requester's copyrighted works for training; (2) the subpoena's purpose is to obtain training material copies or identifying records to determine whether the developer used those works; and (3) the copies or records will be used only to protect the requester's copyright interests.
(d) CONTENTS OF SUBPOENA.—A subpoena issued pursuant to a request under subsection (b) shall authorize and order the developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) receiving the subpoena to expeditiously disclose to the legal or beneficial owner or authorized person all records described in that subsection.
(e) BASIS FOR GRANTING SUBPOENA.—If a proposed subpoena described in subsection (c)(1) is in proper form, and the accompanying declaration described in subsection (c)(2) is properly executed, the clerk shall expeditiously issue and sign the proposed subpoena and return it to the requester for delivery to the developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3).
These subsections define what the subpoena authorizes and the ministerial standard for its issuance. The subpoena orders the developer to expeditiously disclose all records described in subsection (b). The clerk issues the subpoena without judicial review if the proposed subpoena is in proper form and the sworn declaration is properly executed — this is a purely clerical function, not a contested proceeding.
(f) 1 ACTIONS OF DEVELOPERdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) RECEIVING SUBPOENA.—Upon receipt of a subpoena issued under subsection (e), a developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) shall expeditiously disclose to the legal or beneficial owner or authorized person the copies or records requested by the subpoena.
This subsection imposes the bill's core affirmative obligation: upon receipt of a validly issued subpoena, a developer must expeditiously disclose to the copyright owner (or authorized representative) the copies or records identified in the subpoena. This is the only provision in the bill that imposes a new compliance duty on an AI developer — all other subsections either define terms, establish procedural mechanics, or create consequences for noncompliance.
(g) 2 DUTY OF CONFIDENTIALITY.—A legal or beneficial owner or authorized person who receives copies or records from a developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) under this section may not disclose the copies or records to any other person without proper authorization or consent.
This subsection imposes a confidentiality obligation on the copyright owner or authorized person who receives training data copies or records. The recipient may not disclose the materials to any other person without proper authorization or consent. This protects developers' proprietary training data information from being shared beyond the requesting copyright owner.
(h) RULES APPLICABLE TO SUBPOENA.—Unless otherwise provided by this section or by applicable rules of the court, the procedure for issuance and delivery of a subpoena issued under subsection (e), and the remedies for noncompliance with the subpoena, shall be governed to the greatest extent practicable by the provisions of the Federal Rules of Civil Procedure governing the issuance, service, and enforcement of a subpoena duces tecum.
This subsection provides that the Federal Rules of Civil Procedure governing subpoena duces tecum apply to the issuance, service, and enforcement of subpoenas under this section, to the greatest extent practicable. This incorporates existing contempt and enforcement mechanisms without creating new ones.
(i) 3 REBUTTABLE PRESUMPTION.—If a developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) fails to comply with a subpoena issued under subsection (e), that failure shall provide a rebuttable presumption that the developerdeveloperThe term 'developer'— (A) means a person or State or local government agency that— (i) designs, codes, produces, owns, or substantially modifies a generative artificial intelligence model for use by— (I) the person or State or local government agency; or (II) a third party; and (ii) engages in or supervises, including as a third party training dataset curator— (I) the curation of the training dataset of the artificial intelligence model; or (II) the use of the training dataset to train the artificial intelligence model; and (B) does not include a noncommercial end user of a generative artificial intelligence model.17 U.S.C. § 514(a)(3) made copies of the copyrighted work.
This subsection creates a significant evidentiary consequence for developers who fail to comply with a Section 514 subpoena: a rebuttable presumption that the developer made copies of the copyrighted work. This presumption would shift the burden of proof in any subsequent copyright infringement litigation, giving the subpoena mechanism real teeth even though the bill itself does not create a cause of action.
(j)(1)–(2) SANCTIONS FOR BAD FAITH SUBPOENA REQUEST.— (1) MOTION.—If the legal or beneficial owner of an exclusive right under a copyright, or a person authorized to act on the owner's behalf, requests a subpoena under subsection (b) in bad faith, the court that issued the subpoena, upon motion of the recipient of the subpoena, may impose sanctions on the legal or beneficial owner or authorized person. (2) IMPLEMENTATION.—Rule 11(c) of the Federal Rules of Civil Procedure shall apply to sanctions imposed under this subsection in the same manner as that rule applies to sanctions imposed for a violation of rule 11(b) of those Rules.
This subsection provides a safeguard against abuse of the subpoena mechanism. If a copyright owner requests a subpoena in bad faith, the developer may move the issuing court to impose sanctions under the standard of FRCP Rule 11(c). This protects developers from harassment through frivolous or vexatious subpoena requests.
(k) EFFECTIVE DATE.—This section shall take effect on the date of enactment of this section.
This subsection provides that Section 514 takes effect on the date of enactment — i.e., immediately upon the bill becoming law, with no delayed operative date.