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) "Artificial intelligenceArtificial intelligence"Artificial intelligence" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.R.I. Gen. Laws § 9-1-55(a)(1)" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.
(a)(2) "Covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2)" means either of the following: (i) An artificial intelligenceArtificial intelligence"Artificial intelligence" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.R.I. Gen. Laws § 9-1-55(a)(1) model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developerDeveloper"Developer" means a person that performs the initial training of a covered model either by training a model using a sufficient quantity of computing power and cost, or by fine-tuning an existing covered model or covered model derivative using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.R.I. Gen. Laws § 9-1-55(a)(5); or (ii) An artificial intelligenceArtificial intelligence"Artificial intelligence" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.R.I. Gen. Laws § 9-1-55(a)(1) model created by fine-tuningFine-tuning"Fine-tuning" means adjusting the model weights of a trained covered model or covered model derivative by exposing it to additional data.R.I. Gen. Laws § 9-1-55(a)(6) a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developerDeveloper"Developer" means a person that performs the initial training of a covered model either by training a model using a sufficient quantity of computing power and cost, or by fine-tuning an existing covered model or covered model derivative using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.R.I. Gen. Laws § 9-1-55(a)(5), exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuningFine-tuning"Fine-tuning" means adjusting the model weights of a trained covered model or covered model derivative by exposing it to additional data.R.I. Gen. Laws § 9-1-55(a)(6).
(a)(3)–(4) "Covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4)" means any of the following: (i) An unmodified copy of a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2); (ii) A copy of a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) that has been subjected to post-training modifications unrelated to fine-tuningFine-tuning"Fine-tuning" means adjusting the model weights of a trained covered model or covered model derivative by exposing it to additional data.R.I. Gen. Laws § 9-1-55(a)(6); (iii) A copy of a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developerDeveloper"Developer" means a person that performs the initial training of a covered model either by training a model using a sufficient quantity of computing power and cost, or by fine-tuning an existing covered model or covered model derivative using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.R.I. Gen. Laws § 9-1-55(a)(5), does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuningFine-tuning"Fine-tuning" means adjusting the model weights of a trained covered model or covered model derivative by exposing it to additional data.R.I. Gen. Laws § 9-1-55(a)(6); or (4) A copy of a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) that has been combined with other software.
(a)(5) "DeveloperDeveloper"Developer" means a person that performs the initial training of a covered model either by training a model using a sufficient quantity of computing power and cost, or by fine-tuning an existing covered model or covered model derivative using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.R.I. Gen. Laws § 9-1-55(a)(5)" means a person that performs the initial training of a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) either by training a model using a sufficient quantity of computing power and cost, or by fine-tuningFine-tuning"Fine-tuning" means adjusting the model weights of a trained covered model or covered model derivative by exposing it to additional data.R.I. Gen. Laws § 9-1-55(a)(6) an existing covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.
(a)(6) "Fine-tuningFine-tuning"Fine-tuning" means adjusting the model weights of a trained covered model or covered model derivative by exposing it to additional data.R.I. Gen. Laws § 9-1-55(a)(6)" means adjusting the model weights of a trained covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) by exposing it to additional data.
Subsection (a) defines the key terms for the strict liability framework. Covered model is defined by a dual threshold — compute power exceeding 1026 FLOP and cost exceeding $100 million — which targets only frontier-scale models. Fine-tuned models reach covered status only if the fine-tuning itself exceeds 3×1025 FLOP and $10 million. Covered model derivative sweeps in unmodified copies, post-training modifications, lower-cost fine-tunes, and software combinations — ensuring that downstream copies of a covered model remain within the liability regime even if the downstream party is not itself a developer. Developer is limited to the person performing the initial training or the high-threshold fine-tuning, excluding deployers, users, and lower-cost fine-tuners from strict liability.
(b) 1 Except with respect to any causes of action for defamation, developersDeveloper"Developer" means a person that performs the initial training of a covered model either by training a model using a sufficient quantity of computing power and cost, or by fine-tuning an existing covered model or covered model derivative using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.R.I. Gen. Laws § 9-1-55(a)(5) of covered modelsCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativesCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) shall be strictly liable, regardless of the degree of care they exercised, for all injuries to a non-user of the covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) that satisfy the actual harm element of an ordinary negligence claim if: (1) Those injuries are factually and proximately caused by a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) that engages in conduct that, if undertaken by an adult human of sound mind, would satisfy the elements of negligence or any intentional tort or crime; and (2) That conduct was neither intended nor could have been reasonably anticipated by: (i) The user of the model; or (ii) Any intermediary that fine-tuned, scaffolded, or otherwise modified the model.
Subsection (b) is the operative liability provision. It imposes strict liability on developers of covered models and covered model derivatives for injuries to non-users — i.e., third parties who did not themselves interact with the model. Strict liability applies regardless of the degree of care the developer exercised, but requires the plaintiff to satisfy the actual-harm element of an ordinary negligence claim. The AI's conduct must be factually and proximately caused by the model engaging in conduct that would satisfy the elements of negligence, an intentional tort, or a crime if undertaken by a human of sound mind, and the harmful conduct must not have been intended or reasonably anticipated by the user or any intermediary. Defamation claims are expressly excluded from strict liability treatment.
(c)(1) 2 For the purposes of this section, any torts for which the mental state of the alleged tortfeasor is relevant to elements of the tort, there shall be a rebuttable presumption that the artificial intelligenceArtificial intelligence"Artificial intelligence" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.R.I. Gen. Laws § 9-1-55(a)(1) system satisfies the relevant mental state if the finder of fact determines by a preponderance of the evidence that, if a natural person under similar circumstances to the artificial intelligenceArtificial intelligence"Artificial intelligence" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.R.I. Gen. Laws § 9-1-55(a)(1) system took actions similar to those taken by the artificial intelligenceArtificial intelligence"Artificial intelligence" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.R.I. Gen. Laws § 9-1-55(a)(1) system, that natural person would have acted with the relevant mental state.
(c)(2) 2 Unless the court determines that the presumption established in subsection (c)(1) of this section is not applicable, if the party against whom the presumption is invoked presents evidence tending to rebut the presumption established in subsection (c)(1) of this section, the court shall instruct the finder of fact to find that the presumed facts exist unless the finder of fact is persuaded that the presumed facts do not exist.
(c)(3) 2 For the purposes of this section, it shall not be a defense that artificial intelligenceArtificial intelligence"Artificial intelligence" means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.R.I. Gen. Laws § 9-1-55(a)(1) systems are incapable of having mental states.
Subsection (c) addresses the thorny doctrinal problem of attributing mental states to AI systems — a prerequisite for intentional tort and many negligence claims. It creates a rebuttable presumption that the AI satisfies the relevant mental-state element if a natural person in similar circumstances taking similar actions would have acted with that mental state. The presumption is assessed by the finder of fact under a preponderance standard. Even if the opposing party introduces rebuttal evidence, the court must instruct the jury to find the presumed facts exist unless persuaded otherwise — making this a strong, Morgan-style presumption. Subsection (c)(3) expressly forecloses the defense that AI systems are incapable of having mental states.
(d)(1) 3 It shall be an affirmative defense to strict liability if the developerDeveloper"Developer" means a person that performs the initial training of a covered model either by training a model using a sufficient quantity of computing power and cost, or by fine-tuning an existing covered model or covered model derivative using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.R.I. Gen. Laws § 9-1-55(a)(5) establishes that the covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) satisfied the standard of care applicable to humans who perform the same function that the covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) was engaged in performing when its conduct allegedly caused the plaintiff's injury.
(d)(2) 3 It shall be an affirmative defense to strict liability if the developerDeveloper"Developer" means a person that performs the initial training of a covered model either by training a model using a sufficient quantity of computing power and cost, or by fine-tuning an existing covered model or covered model derivative using a quantity of computing power and cost greater than the amount specified in subsection (a)(2) of this section.R.I. Gen. Laws § 9-1-55(a)(5) establishes that the injuries to a non-user as described in this section were a result of a capabilities failure, in which a covered modelCovered model"Covered model" means either of the following: (i) An artificial intelligence model trained using a quantity of computing power greater than 10^26 integer or floating-point operations, the cost of which exceeds one hundred million dollars ($100,000,000) when calculated using the average market prices of cloud compute at the start of training as reasonably assessed by the developer; or (ii) An artificial intelligence model created by fine-tuning a covered model using a quantity of computing power equal to or greater than three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, exceeds ten million dollars ($10,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning.R.I. Gen. Laws § 9-1-55(a)(2) or covered model derivativeCovered model derivative"Covered model derivative" means any of the following: (i) An unmodified copy of a covered model; (ii) A copy of a covered model that has been subjected to post-training modifications unrelated to fine-tuning; (iii) A copy of a covered model that has been fine-tuned using a quantity of computing power not exceeding three (3) times 10^25 integer or floating-point operations, the cost of which, as reasonably assessed by the developer, does not exceed one million dollars ($1,000,000) if calculated using the average market price of cloud compute at the start of fine-tuning; or (4) A copy of a covered model that has been combined with other software.R.I. Gen. Laws § 9-1-55(a)(3)–(4) falls short of performing the intended or reasonably anticipated performance of the user, but the conduct of the system would not satisfy the elements of negligence or any intentional tort or crime if engaged in by an adult human of sound mind.
Subsection (d) provides two affirmative defenses that developers may raise against strict liability. The first defense applies where the developer can show the model satisfied the standard of care applicable to humans performing the same function the model was engaged in when the injury occurred — effectively importing professional-negligence standards into the AI context. The second defense applies to capabilities failures: where the model simply fell short of its intended performance but did not engage in conduct that would constitute negligence or a tort if performed by a human. Together, these defenses narrow the strict liability regime to injuries caused by genuinely tortious autonomous model conduct, not mere performance shortfalls.
This act shall take effect upon passage.
The act takes effect upon passage — there is no delayed operative date or phase-in period.