,,ii #aaith ,,congress #b;d ,session ,s4 #ddgf ,,in ,,the ,,senate ,,of ,,the ,,united ,,states ,april #cj1 #bjbf,mr4 ,warner 7for himself and ,mr4 ,budd7 introduced the following bill2 which was read twice and referred to the ,committee on ,health1 ,education1 ,labor1 and ,pensions ,a ,,bill ,to provide for voluntary disclosure by developers and users of artificial intelligence regarding workforce data and for reporting by the ,secretary of ,labor regarding the workforce data1 and for other purposes4 #a4 ,short title ,this ,act may be cited as the ,workforce ,transparency ,act of #bjbf4 #b4 ,findings and purposes 7a7 ,findings ,congress finds the following3 7#a7 ,artificial intelligence systems are already1 as of the date of enactment of this ,act1 widely used by workers1 employers1 and consumers across sectors of the ,united ,states economy4 7#b7 ,the most immediate labor-market effects of artificial intelligence are currently occurring at the task1 skill1 and workflow level1 rather than through the wholesale elimination of occupations4 7#c7 ,research shows that artificial intelligence is also improving workforce productivity by accelerating routine cognitive tasks1 enabling workers to focus on higher-value functions1 and contributing positively to aggregate economic growth1 even as adoption and impacts vary by sector and over time4 ,artificial intelligence also has the potential to create new roles1 consistent with previous patterns in which technological change has led to significant changes in the types of jobs people do over time4 7#d7 ,policymakers1 workers1 educators1 and employers lack timely and standardized information about how artificial intelligence systems are being used and how such use affects productivity1 skills1 and workforce outcomes4 7#e7 ,aggregated1 privacy-preserving transparency1 including information voluntarily shared by a covered ,,ai system provider and complementary data published by the ,federal ,government1 can meaningfully improve workforce development1 education policy1 and economic planning without revealing proprietary information1 trade secrets1 individual-level data1 or other information that would undermine individual privacy4 7#f7 ,a ,federal transparency framework that encourages participation and supports modernization of government data systems can establish a consistent national baseline1 reduce fragmentation across ,states1 and provide a level playing field among covered ,,ai system providers4 7b7 ,purposes ,the purposes of this ,act are to-- 7#a7 establish a ,federal framework for aggregated1 de-identified transparency or data-sharing regarding use of artificial intelligence systems relevant to workforce impacts2 7#b7 support evidence-based workforce1 education1 and economic policymaking through both voluntary participation by a covered ,,ai system provider and improved ,federal labor-market data2 7#c7 protect individual privacy1 confidential business information1 and competition2 and 7#d7 promote public trust through consistent1 responsible1 and privacy-preserving data-sharing4 #c4 ,definitions ,in this ,act3 7#a7 ,aggregated workforce transparency data ,the term aggregated workforce transparency data means statistical information that-- 7,a7 is aggregated across users1 accounts1 interactions1 or customers2 7,b7 does not identify or reasonably permit the identification of a particular person1 household1 or employer2 and 7,c7 does not disclose-- 7i7 proprietary model weights1 training data1 source code1 or system architecture2 or 7ii7 information that a covered ,,ai system provider has contractually defined as confidential or not subject to disclosure4 7#b7 ,artificial intelligence ,the term artificial intelligence has the meaning given the term in section #ejjb of the ,national ,artificial ,intelligence ,initiative ,act of #bjbj 7#ae ,u4,s4,c4 #idja74 7#c7 ,covered ,,ai system ,the term covered ,,ai system means an artificial intelligence system made available to consumers or businesses other than such a system that is-- 7,a7 designed and marketed by a person primarily for internal business use by the person2 7,b7 used solely for academic or non-commercial research2 or 7,c7 designed for a narrow1 discrete technical function without general-purpose capabilities4 7#d7 ,covered ,,ai system provider ,the term covered ,,ai system provider means a person that develops1 deploys1 or makes available a covered ,,ai system4 7#e7 ,enterprise customer ,the term enterprise customer means a business1 governmental1 or institutional customer that accesses a covered ,,ai system pursuant to a negotiated contract governing data use1 confidentiality1 or disclosure4 7#f7 ,participating entity ,the term participating entity means a covered ,,ai system provider or enterprise customer that elects to submit data in accordance with section #d7;a74 7#g7 ,secretary ,the term ,secretary means the ,secretary of ,labor4 #d4 ,workforce transparency reporting 7a7 ,in general ,a participating entity may submit to the ,secretary1 in accordance with this ,act and any regulations promulgated by the ,secretary under this ,act1 aggregated workforce transparency data1 including 7as reasonably feasible7 such data regarding the following3 7#a7 ,task or activity level use ,aggregated distribution of interactions across broad task or activity categories1 such as writing1 coding1 research1 analysis1 translation1 or job-related assistance 7as established and periodically updated by the ,secretary through notice and comment rulemaking under section #eec of title #e1 ,united ,states ,code74 7#b7 ,geographic distribution ,aggregated distribution of interactions across broad task or activity categories1 by ,state or by metropolitan statistical area or comparable geographic unit4 7#c7 ,age ranges ,aggregated distribution of interactions across broad task or activity categories1 by adult age ranges1 as reasonably feasible and consistent with privacy safeguards or laws4 7#d7 ,temporal trends ,changes in usage patterns over time based on task or activity category4 7b7 ,reporting windows ,the ,secretary shall establish a system for submissions under subsection 7a7 that allows for such submissions by a participating entity only for aggregated workforce transparency data concerning information that is not more recent than #ia days before the date of the submission1 including1 for purposes of the first submission by the participating entity1 such data with respect to the #b-year period preceding the date on which the participating entity first elects to make such a submission4 7c7 ,rule of construction ,this ,act shall not be construed to require a covered ,,ai system provider or enterprise customer to elect to submit or otherwise be required to submit aggregated workforce transparency data under this section4 #e4 ,privacy1 security1 and confidentiality safeguards 7a7 ,prohibited disclosures ,a participating entity may not submit under section #d7;a7-- 7#a7 personal data or information linked or reasonably linkable to an individual2 7#b7 employer-specific1 customer-specific1 or individual-specific performance data2 7#c7 trade secrets1 proprietary algorithms1 model weights1 training datasets1 or source code2 7#d7 information that a participating entity is contractually prohibited from disclosing pursuant to agreements or terms with consumers1 users1 or enterprise1 governmental1 or institutional customers or that the participating entity otherwise commits to protecting from disclosure in policies or representations1 including restrictions or commitments on data use1 aggregation1 or secondary disclosure2 or 7#e7 data in violation of a contract or other terms with an enterprise customer1 consumer1 or user1 including contractual restrictions on data use1 aggregation1 disclosure1 or secondary analysis4 7b7 ,compliance with applicable laws ,in making submissions under section #d7;a71 a participating entity shall exclude or further aggregate data as necessary to comply with other applicable laws4 7c7 ,compliance with contract terms ,this ,act shall not be construed to obviate the terms1 including privacy or data protections1 in any contract between a participating entity and its customers1 users1 or other consumers4 7d7 ,safeguards ,in preparing data for purposes of making submissions under section #d7;a7 and making such submissions1 a participating entity shall apply aggregation1 anonymization1 and de-identification techniques consistent with guidance issued under section #f7;c74 7e7 ,freedom of information act ,any aggregated workforce transparency data submitted under section #d7;a7 shall be exempt from the disclosure requirements under section #eeb of title #e1 ,united ,states ,code 7commonly known as the ,freedom of ,information ,act74 #f4 ,reporting and regulations 7a7 ,lead agency ,the ,secretary of ,labor1 acting through the ,commissioner of ,labor ,statistics and in coordination with the ,secretary of ,commerce acting through the ,director of the ,bureau of the ,census1 shall administer this ,act4 7b7 ,data reporting 7#a7 ,public database ,the ,secretary shall establish and maintain a publicly available1 online database containing aggregated workforce transparency data derived from submissions under section #d7;a74 7#b7 ,annual report to ,congress ,the ,secretary shall annually submit a report to ,congress summarizing aggregated workforce transparency data derived from submissions under such section4 7#c7 ,requirements for data ,in carrying out this subsection1 the ,secretary shall aggregate data from submissions by participating entities under section #d7;a7 and ensure that any published dataset1 analysis1 or report does not attribute any submitted data to any specific participating entity or any identifiable person4 7c7 ,guidance ,not later than #ahj days after the date of enactment of this ,act1 the ,secretary shall issue non-binding guidance regarding-- 7#a7 standardized reporting formats and taxonomies for purposes of submissions under section #d7;a72 7#b7 acceptable aggregation1 anonymization1 and privacy-preserving methods for purposes of submissions under such section2 7#c7 procedures to minimize reporting burden for purposes of such submissions1 including alignment with existing ,federal data-collection efforts2 and 7#d7 definitions and classifications for broad task or activity categories used in workforce-related artificial intelligence usage reporting4 7d7 ,rulemaking 7#a7 ,in general ,not later than #ahj days after the date of enactment of this ,act1 the ,secretary shall issue regulations necessary to carry out this ,act4 7#b7 ,requirements ,the regulations issued under paragraph 7#a7 shall clarify the extent to which a participating entity shall aggregate1 anonymize1 or otherwise ensure privacy safeguards for aggregated workforce transparency data submitted under section #d7;a74 7e7 ,working group ,the ,secretary1 acting through the ,commissioner of ,labor ,statistics and in coordination with other relevant agencies1 shall establish a working group to develop a process to publish or enhance public labor-market data that combines age and occupational information1 in order to improve understanding of how artificial intelligence affects entry-level workers and early-career professionals4 #g4 ,prohibition regarding adverse inferences ,a ,federal agency may not draw an adverse inference against any covered ,,ai system provider or enterprise customer that elects not to make submissions under section #d7;a74 #h4 ,enforcement ,the ,secretary may seek injunctive relief in any court of competent jurisdiction against a participating entity with respect to aggregated workforce transparency data submitted under section #d7;a7 that the ,secretary determines is a knowing and willful misrepresentation by the entity of the data4 #i4 ,severability ,if any provision of this ,act or the application of such provision to any person or circumstance is held to be unconstitutional1 the remainder of this ,act and the application of the provision to any other person or circumstance1 shall not be affected4