,i #aaith ,,congress #b;d ,session ,h4 ,r4 #ifai ,,in ,,the ,,house ,,of ,,representatives ,july #i1 #bjbf,mrs4 ,foushee 7for herself and ,mr4 ,casar7 introduced the following bill2 which was referred to the ,committee on ,energy and ,commerce ,a ,,bill ,to require artificial intelligence chatbot providers to provide data privacy and security1 and for other purposes4 #a4 ,short title ,this ,act may be cited as the ,people-,first ,chatbot ,act4 #b4 ,prohibitions and requirements 7a7 ,data privacy and security 7#a7 ,prohibitions ,an artificial intelligence chatbot provider may not do the following3 7,a7 ,process personal data other than input data to inform chatbot outputs unless the processing of personal data is necessary to fulfill an express request made by a user and that user has provided affirmative consent4 7,b7 ,process the chat log of a user as follows3 7i7 ,to determine whether to display an advertisement for a product or service to the user4 7ii7 ,to determine a product1 service1 or category of product or service to advertise to the user4 7iii7 ,to customize an advertisement or how an advertisement is presented to the user4 7,c7 ,process the chat log or personal data of a user 7unless the personal data is the age of the user1 but only to implement the policies of the artificial intelligence chatbot provider1 regarding harmful design features for minors7 as follows3 7i7 ,if the artificial intelligence chatbot provider knows or should know1 based on knowledge fairly implied on the basis of objective circumstances1 that the user is under the age of #ah1 without the affirmative consent of the parent or legal guardian of that user4 7ii7 ,for training purposes1 if the artificial intelligence chatbot provider knows or should know1 based on knowledge fairly implied on the basis of objective circumstances1 that a user is under #ah years of age4 7iii7 ,for training purposes1 of a user over #ah years of age1 unless the artificial intelligence chatbot provider first obtains affirmative consent4 7,d7 ,use any classification or designation of the personality or behavioral characteristic of a user created through profiling beyond what is necessary to fulfill an express request made by a user4 7,e7 ,sell the chat log of a user4 7,f7 ,retain the chat log of a user for longer than #e years1 unless retention is necessary to comply with this ,act or otherwise required by law4 7,g7 ,discriminate or retaliate against any user1 including by denying products or services1 charging different prices or rates for products or services1 or providing lower quality products or services to the user1 for refusing to consent to the use of chat logs or personal data for training purposes4 7#b7 ,right of access 7,a7 ,access to chat log ,an artificial intelligence chatbot provider that retains the chatlog of a user shall retain such chat log in a portable1 readily usable1 downloadable1 and human and machine-readable format1 and shall provide such chat log to the user upon request by the user4 7,b7 ,deletion of chat logs and personal data ,an artificial intelligence chatbot provider that retains the chat log or personal data of a user shall delete such chat log or personal data upon request by the user4 7,c7 ,discrimination and retaliation prohibited ,an artificial intelligence chatbot provider may not discriminate or retaliate against any user1 including by denying products or services1 charging different prices or rates for products or services1 or providing lower quality products or services to the user1 for accessing the chat log of the user4 7#c7 ,data security program ,not later than #ab months after the date of the enactment of this ,act1 an artificial intelligence chatbot provider shall develop1 implement1 and maintain a comprehensive1 written1 data security program that contains administrative1 technical1 and physical safeguards that are proportionate to the volume and nature of the personal data and chat logs maintained by the artificial intelligence chatbot provider1 a summary of which shall be made publicly available on the website of the artificial intelligence chatbot provider4 7#d7 ,prohibited production and access ,a ,government entity may not compel the production of or access to input data or chat logs from an artificial intelligence chatbot provider1 except as provided by a warrant issued by a court under section #beah of title #ah1 ,united ,states ,code4 7b7 ,transparency for users 7#a7 ,prohibition against misleading information ,an artificial intelligence chatbot provider may not-- 7,a7 represent that the input data or chat log of a user is confidential2 or 7,b7 use any term1 letter1 or phrase in the advertising1 interface1 or output of an artificial intelligence chatbot that indicates or implies that any output data is being provided by1 endorsed by1 or equivalent to those provided by the following3 7i7 ,a licensed healthcare professional or a licensed therapist4 7ii7 ,a licensed legal professional4 7iii7 ,a licensed accounting professional4 7iv7 ,a certified financial fiduciary or planner4 7#b7 ,,ai chatbot notice required ,an artificial intelligence chatbot provider shall provide a clear1 conspicuous1 and explicit notice to a user that the user is interacting with an artificial intelligence chatbot rather than a human before the artificial intelligence chatbot generates any output1 every hour thereafter1 and each time a user prompts the artificial intelligence chatbot about whether the artificial intelligence chatbot is a real person as follows3 7,a7 ,the disclosure of the notice shall be in the same language as the one in which the user interacts with the artificial intelligence chatbot1 in a font size easily readable by an average user1 and not smaller than the largest font size of other text appearing on the interface on which the artificial intelligence chatbot is provided4 7,b7 ,the notice shall be accessible to users with disabilities4 7,c7 ,the notice shall comply with regulations promulgated by the ,commission4 7c7 ,safety by design3 assessments and transparency requirements 7#a7 ,monthly risk assessment of ,,ai chatbot required ,not less frequently than monthly1 an artificial intelligence chatbot provider shall assess the artificial intelligence chatbot for the risk of any covered harm1 emotional dependence1 or compulsive usage1 according to metrics set forth in rules promulgated by the ,commission4 7#b7 ,public availability of ,,ai chatbot information ,not less frequently than quarterly1 an artificial intelligence chatbot provider shall make information concerning the risk assessment required under paragraph 7#a7 publicly available on the website of the provider1 in accordance with rules promulgated by the ,commission4 7#c7 ,prohibition of harmful design features for minors ,if an artificial intelligence chatbot provider knows or should know1 based on knowledge fairly implied on the basis of objective circumstances1 that a user is under the age of #ah1 the artificial intelligence chatbot provider shall disable for that user any feature or setting that creates an unreasonable risk of the user suffering a covered harm1 emotional dependence1 or compulsive usage1 in accordance with rules promulgated by the ,commission4 7d7 ,required disclosure by business entities engaged in customer service communications of use of an artificial intelligence chatbot for customer service ,a business entity that initiates or receives a customer service communication and uses an artificial intelligence chatbot for customer service communication shall1 at the beginning of each customer service communication so initiated or received1 disclose1 in accordance with rules promulgated by the ,commission-- 7#a7 that a nonhuman1 artificial intelligence1 or machine is being used for customer service2 and 7#b7 that the consumer may request1 and upon such request1 be immediately transferred to a human operator who is physically located in the ,united ,states1 including1 if possible1 by voice command 7such as by saying the word agent74 #c4 ,rulemaking 7a7 ,regulations required ,not later than #ab months after the date of the enactment of this ,act1 the ,commission shall promulgate regulations that include the following3 7#a7 ,a description of the form and content of the disclosures required under section #b7;b74 7#b7 ,an example template for the disclosures required under section #b7;b74 7#c7 ,a description of the metrics that each artificial intelligence chatbot provider is required to use to assess and publish any risk of a covered harm1 emotional dependence1 or compulsive usage under section #b7;c77#c74 7#d7 ,rules that prohibit the use of artificial intelligence chatbot design features or settings that create an unreasonable risk of causing a covered harm1 emotional dependence1 or compulsive usage for users under the age of #ah1 as described under section #b7;c77#c74 7#e7 ,an identification and description of categories of information that each artificial intelligence chatbot provider is required to make publicly available about the artificial intelligence chatbots of the provider under section #b7;b74 7#f7 ,a description and example of the disclosure required by each business entity that initiates or receives a customer service communication and uses an artificial intelligence chatbot for customer service communication1 as described under section #b7;d74 7b7 ,general rulemaking authority ,the ,commission may promulgate any other regulation necessary to implement this ,act4 #d4 ,enforcement 7a7 ,enforcement by ,federal ,trade ,commission 7#a7 ,unfair or deceptive acts or practices ,a violation of section #b or a regulation promulgated under such section shall be treated as a violation of a regulation under section #ah7;a77#a77;,b7 of the ,federal ,trade ,commission ,act 7#ae ,u4,s4,c4 #eg;a7a77#a77;,b77 regarding unfair or deceptive acts or practices4 7#b7 ,powers of ,commission ,the ,federal ,trade ,commission shall enforce section #b and any regulation promulgated under such section in the same manner1 by the same means1 and with the same jurisdiction1 powers1 and duties as though all applicable terms and provisions of the ,federal ,trade ,commission ,act 7#ae ,u4,s4,c4 #da et seq47 were incorporated into and made a part of this section4 ,any person who violates such section or regulation shall be subject to the penalties and entitled to the privileges and immunities provided in the ,federal ,trade ,commission ,act4 7b7 ,actions by ,states 7#a7 ,in general ,in any case in which the attorney general of a ,state1 or an official or agency of a ,state1 has reason to believe that an interest of the residents of such ,state has been or is threatened or adversely affected by an act or practice in violation of section #b or a regulation promulgated under such section1 the ,state1 as parens patriae1 may bring a civil action on behalf of the residents of the ,state in an appropriate ,state court or appropriate district court of the ,united ,states to-- 7,a7 enjoin such act or practice2 7,b7 enforce compliance with such section or regulation2 7,c7 obtain damages1 restitution1 or other compensation on behalf of residents of the ,state2 7,d7 obtain reasonable attorney's fees and other litigation costs reasonably incurred2 or 7,e7 obtain such other legal and equitable relief as the court may consider to be appropriate4 7#b7 ,notice ,before filing an action under this subsection1 the attorney general1 official1 or agency of the ,state involved shall provide to the ,federal ,trade ,commission a written notice of such action and a copy of the complaint for such action4 ,if the attorney general1 official1 or agency determines that it is not feasible to provide the notice described in this paragraph before the filing of the action1 the attorney general1 official1 or agency shall provide written notice of the action and a copy of the complaint to the ,federal ,trade ,commission immediately upon the filing of the action4 7#c7 ,authority of ,federal ,trade ,commission 7,a7 ,in general ,on receiving notice under paragraph 7#b7 of an action under this subsection1 the ,federal ,trade ,commission shall have the right-- 7i7 to intervene in the action2 7ii7 upon so intervening1 to be heard on all matters arising therein2 and 7iii7 to file petitions for appeal4 7,b7 ,limitation on ,state action while ,federal action is pending ,if the ,federal ,trade ,commission or the ,attorney ,general of the ,united ,states has instituted a civil action for violation of section #b or a regulation promulgated under such section 7referred to in this subparagraph as the ,federal action71 no ,state attorney general1 official1 or agency may bring an action under this subsection during the pendency of the ,federal action against any defendant named in the complaint in the ,federal action for any violation of such section or regulation alleged in such complaint4 7#d7 ,rule of construction ,for purposes of bringing a civil action under this subsection1 nothing in this ,act shall be construed to prevent an attorney general1 official1 or agency of a ,state from exercising the powers conferred on the attorney general1 official1 or agency by the laws of such ,state to conduct investigations1 administer oaths and affirmations1 or compel the attendance of witnesses or the production of documentary and other evidence4 7c7 ,private right of action 7#a7 ,in general ,a person injured by an act or practice in violation of section #b or a regulation promulgated under such section may bring in an appropriate ,state court or an appropriate district court of the ,united ,states any of the following3 7,a7 ,an action to enjoin the violation4 7,b7 ,an action to recover actual damages resulting from the violation1 or to receive-- 7i7 up to #aj1jjj per violation1 for a violation of subsections 7a71 7c77#a71 and 7c77#b7 of section #b1 whichever is greater2 and 7ii7 up to #aj1jjj in total for all violations of section #b7;b71 whichever is greater4 7iii7 for a violation of-- 7,i7 section #b7;c77#c71 actual damages and statutory damages of not less than #ej1jjj4 7,,ii7 section #b7;c77#c7 that results in a covered harm1 actual damages and statutory damages of not less than #bej1jjj4 7,,iii7 section #b7;c77#c7 that results in emotional dependence or compulsive usage1 actual damages and statutory damages of not less than #ajj1jjj4 7#b7 ,willful or knowing violations ,if the court finds that the defendant acted willfully or knowingly in committing a violation described in paragraph 7#a71 the court may1 in its discretion1 increase the amount of the award to an amount equal to not more than #e times the amount available under paragraph 7#a77;,b74 7#c7 ,costs and attorney's fees ,the court shall award to a prevailing plaintiff in an action under this subsection the costs of such action and reasonable attorney's fees1 as determined by the court4 7#d7 ,limitation ,an action may be commenced under this subsection not later than #b years after the date on which the person first discovered or had a reasonable opportunity to discover the violation4 7#e7 ,nonexclusive remedy ,the remedy provided by this subsection shall be in addition to any other remedies available to the person4 7d7 ,liability for injury ,a user of an artificial intelligence chatbot who suffers an injury in fact caused through the use of the artificial intelligence chatbot may bring an action against the artificial intelligence chatbot provider in an appropriate ,state court or an appropriate district court of the ,united ,states to recover actual damages from the artificial intelligence chatbot provider1 even if-- 7#a7 the artificial intelligence chatbot provider exercised all reasonable care in the design and distribution of the artificial intelligence chatbot2 or 7#b7 the artificial intelligence chatbot provider did not directly distribute the artificial intelligence chatbot to the user or otherwise enter into a contractual relationship with the user4 #e4 ,definitions ,in this ,act3 7#a7 ,advertisement ,the term advertisement means any written or oral statement1 illustration1 or depiction that promotes the sale or use of a good or service or is designed to increase interest in a brand1 good1 or service in which the statement1 illustration1 or depiction is displayed in exchange for monetary or other valuable consideration1 including access to data1 between the artificial intelligence chatbot provider and the brand1 good1 or service4 7#b7 ,affirmative consent 7,a7 ,requirements ,the term affirmative consent means a clear affirmative act that signifies the freely given1 specific1 informed1 and unambiguous authorization of a user for an act or practice in response to a specific request from an artificial intelligence chatbot provider if each of the following conditions are met3 7i7 ,the request is provided to the user in a clear and conspicuous standalone disclosure4 7ii7 ,the request includes a written description1 in easy-to-understand language1 of the act or practice for which the consent of the user is sought4 7iii7 ,the request is made in a manner reasonably accessible to and usable by a user with a disability 7as defined in section #c of the ,americans with ,disabilities ,act of #aiij 7#db ,u4,s4,c4 #abajb774 7iv7 ,the request is made available to the user in each language in which the artificial intelligence chatbot provider provides an artificial intelligence chatbot4 7v7 ,the option to refuse to give consent is at least as prominent as the option to give consent1 and the option to refuse to give consent takes the same number of steps or fewer as the option to give consent4 7,b7 ,exclusions ,the term affirmative consent does not include any of the following3 7i7 ,inference of consent from the inaction of the user or the continued use of an artificial intelligence chatbot by the user4 7ii7 ,acceptance of a general or broad terms of use or similar document4 7iii7 ,hovering over1 muting1 pausing1 or closing a given piece of content4 7iv7 ,agreement obtained through the use of a false1 fraudulent1 or materially misleading statement or representation4 7v7 ,agreement obtained through the use of other dark patterns4 7#c7 ,artificial intelligence chatbot 7,a7 ,in general ,the term artificial intelligence chatbot-- 7i7 means any interactive computer service or software application that-- 7,i7 generates responses that are not fully predetermined2 and 7,,ii7 accepts open-ended natural-language or multimodal user input and produces adaptive or context-responsive output2 and 7ii7 does not include an interactive computer service or software application-- 7,i7 the responses of which are limited to contextualized replies2 and 7,,ii7 that is unable to respond on a range of topics outside of a narrow specified purpose4 7,b7 ,inclusions ,the term artificial intelligence chatbot includes a service or application described in subparagraph 7,a7 that1 in a manner that simulates a sustained interpersonal relationship or emotional interaction with the user-- 7i7 exhibits persistent responses that suggest affection or attachment directed toward the user1 or engages in interactions involving emotional disclosures from the user2 or 7ii7 presents at least one persistent identity1 persona1 or character or holds itself out as a sentient being1 fictional character1 or social entity4 7#d7 ,artificial intelligence chatbot provider ,the term artificial intelligence chatbot provider means any person who creates1 distributes 7including to a third party71 or otherwise makes publicly available an artificial intelligence chatbot4 7#e7 ,chat log ,the term chat log means any input data1 output generated by an artificial intelligence chatbot1 or record of the input data or output from user interaction with an artificial intelligence chatbot4 7#f7 ,collect ,the term collect means to create1 buy1 rent1 gather1 obtain1 receive1 access1 or otherwise acquire personal data or input data by any means through the use of an artificial intelligence chatbot by an individual4 7#g7 ,commission ,the term ,commission means the ,federal ,trade ,commission4 7#h7 ,compulsive usage ,the term compulsive usage means a persistent and repetitive use of an ,,ai chatbot that significantly impacts one or more major life activities of a user1 including socializing1 sleeping1 eating1 learning1 reading1 concentrating1 communicating1 or working4 7#i7 ,covered harm ,the term covered harm means death1 a suicide attempt1 or a psychiatric emergency that results in urgent medical treatment4 7#aj7 ,dark pattern ,the term dark pattern means a user interface designed or manipulated with the substantial effect of subverting or impairing user autonomy1 decision-making or choice1 and includes any practice the ,commission refers to as a dark pattern4 7#aa7 ,de-identified data ,the term de-identified data means-- 7,a7 information that cannot reasonably be used to infer or derive the identity of an individual2 7,b7 information that does not identify and is not linked or reasonably linkable to an individual2 or 7,c7 a device that identifies or is linked or reasonably linkable to an individual1 regardless of whether the information is aggregated1 if the artificial intelligence chatbot provider-- 7i7 takes such physical1 administrative1 and technical measures as are necessary to ensure that the information cannot1 at any point1 be used to re-identify any individual or device that identifies or is linked or reasonably linkable to an individual2 7ii7 publicly commits in a clear and conspicuous manner to-- 7,i7 process1 retain1 or transfer the information solely in a de-identified form without any reasonable means for re-identification2 and 7,,ii7 not attempt to re-identify the information with any individual or device that identifies or is linked or reasonably linkable to an individual2 and 7iii7 contractually obligates any entity that receives the information from the artificial intelligence chatbot provider to-- 7,i7 comply with all of the provisions of this subparagraph with respect to the information2 and 7,,ii7 require that such contractual obligation is included in any subsequent instance for which the data may be received4 7#ab7 ,emotional dependence emotional dependence means a behavioral or spoken pattern of the user that indicates the user relies on an artificial intelligence chatbot as a primary source of emotional support or social connection1 such as-- 7,a7 a user expressing that the artificial intelligence chatbot is the primary source of emotional support for the user2 7,b7 a user expressing distress at the prospect of losing access to the artificial intelligence chatbot2 or 7,c7 any pattern of use that suggests the user is substituting the artificial intelligence chatbot for human relationships4 7#ac7 ,input data ,the term input data means information1 including text1 photo1 audio1 video1 or file provided to an artificial intelligence chatbot by a user4 7#ad7 ,model ,the term model means an engineered or machine-based system underlying an artificial intelligence chatbot that can1 for explicit or implicit objectives1 infer from the input it receives how to generate outputs that can influence physical or virtual environments4 7#ae7 ,personal data ,the term personal data-- 7,a7 means any information1 including derived data1 inferences1 or unique identifiers1 that is linked or reasonably linkable1 alone or in combination with other information1 to an identified or identifiable individual or a device that identifies or is linked or reasonably linkable to an individual2 and 7,b7 does not include de-identified data or publicly available information4 7#af7 ,publicly available information ,the term publicly available information means information that has been lawfully made available to the general public from-- 7,a7 ,federal1 ,state or municipal government records1 if the person collects1 processes1 and transfers such information in accordance with any restrictions or terms of use placed on the information by the relevant government entity2 7,b7 widely distributed media2 or 7,c7 a disclosure to the general public as required by ,federal1 ,state1 or local law4 7#ag7 ,publicly available information ,the term publicly available information does not include the following3 7,a7 ,any obscene visual depiction 7as defined in section #adfj of title #ah1 ,united ,states ,code74 7,b7 ,biometric data4 7,c7 ,personal data that is created through the combination of personal data with publicly available information4 7,d7 ,information that is collated and combined to create user profiles on publicly available or subscription-based websites and inferences generated from such information4 7,e7 ,genetic data1 unless otherwise made publicly available by the individual to whom the information pertains4 7,f7 ,information made available by a user on a website or online service made available to all members of the public1 for free or for a fee1 where the user has restricted the information to a specific audience4 7,g7 ,intimate images1 authentic or computer-generated1 known to be nonconsensual4 7#ah7 ,process2 processing ,the terms process and processing mean any operation or set of operations performed1 whether by manual or automated means1 on personal data or input data or on sets of personal data or input data1 such as the use1 storage1 disclosure1 analysis1 deletion1 or modification of such data4 7#ai7 ,profiling ,the term profiling-- 7,a7 means any form of processing performed on input data or personal data to infer1 detect1 classify1 or designate emotional vulnerability or distinct behavioral characteristics of an individual2 and 7,b7 does not include processing of a chat log for purposes of user safety or to otherwise comply with this ,act4 7#bj7 ,sell ,the term sell-- 7,a7 means the exchange of personal data or input data for monetary or other valuable consideration1 or making available such data or use of such data1 by the ,,ai chatbot provider to a third party2 and 7,b7 does not include-- 7i7 the disclosure of personal data or input data to a third party that processes the data on behalf of the artificial intelligence chatbot provider2 7ii7 with the affirmative consent of the user1 the disclosure of personal data or input data in which the user affirmatively directs the artificial intelligence chatbot provider to disclose the data or intentionally uses the artificial intelligence chatbot provider to interact with a third party2 or 7iii7 the disclosure of personal data that the user-- 7,i7 intentionally made available to the general public through a channel of mass media2 and 7,,ii7 did not restrict to a specific audience4 7#ba7 ,training ,the term training-- 7,a7 means the use of input data to adjust or modify a model2 and 7,b7 does not include-- 7i7 testing to identify risks of harm to a user2 7ii7 any adjustment or modification to address any such identified risks of harm2 or 7iii7 any action necessary to comply with this ,act or otherwise required by law4 7#bb7 ,user ,the term user means an individual1 regardless of age4 7#bc7 ,widely distributed media ,the term widely distributed media-- 7,a7 means information that is available to the public1 including information from a telephone book or online directory1 a television1 internet1 or radio program1 the news media1 or an internet site that is available to the public on an unrestricted basis2 and 7,b7 does not include an obscene visual depiction 7as defined in section #adfj of title #ah1 ,united ,states ,code74 #f4 ,rule of construction ,nothing in this ,act or any regulation promulgated under this ,act may be construed to affect any right1 cause of action1 remedy1 presumption1 liability1 or defense available at law or in equity1 including any anti-discrimination1 consumer protection1 labor1 tort1 or civil rights law4 #g4 ,relationship to state laws ,nothing in this ,act or any regulation promulgated under this ,act preempts or otherwise affects any ,state law1 rule1 requirement1 or regulation1 including any right1 cause of action1 remedy1 presumption1 liability1 or defense available at law or in equity1 that is at least as protective of users of artificial intelligence chatbots as provided for in this ,act or any regulation promulgated under this ,act4