,i #aaith ,,congress #b;d ,session ,h4 ,r4 #ajahi ,,in ,,the ,,house ,,of ,,representatives ,august #ca1 #bjbf,ms4 ,jacobs 7for herself1 ,mr4 ,moran1 and ,mr4 ,whitesides7 introduced the following bill2 which was referred to the ,committee on ,armed ,services ,a ,,bill ,to amend title #aj1 ,united ,states ,code1 to direct the ,secretary of ,defense to establish an artificial intelligence incident and vulnerability reporting program1 and for other purposes4 #a4 ,short title ,this ,act may be cited as the ,defense ,,ai ,reliability and ,reporting ,act4 #b4 ,department of ,defense ,,ai incident and vulnerability reporting program ,chapter #aca of title #aj1 ,united ,states ,code1 is amended by inserting after section #bbbd;a the following new section3 #bbbd;b4 ,artificial intelligence incident and vulnerability reporting program 7a7 ,in general ,the ,secretary of ,defense shall establish a centralized ,department-wide program for the reporting1 tracking1 analysis1 and remediation of covered ,,ai incidents and covered ,,ai vulnerabilities arising from the development1 testing1 procurement1 fielding1 or operation of artificial intelligence systems within the ,department of ,defense4 7b7 ,purpose ,the purpose of the program established under subsection 7a7 shall be to-- 7#a7 identify recurring risks1 failure modes1 vulnerabilities1 and systemic weaknesses in artificial intelligence systems1 including risks or failure modes arising from human-machine teaming2 7#b7 support mitigation of significant risks2 and 7#c7 inform testing1 procurement1 cybersecurity1 and deployment decisions to improve the safety1 security1 reliability1 and operational effectiveness of such systems4 7c7 ,requirements for program ,the program shall-- 7#a7 be designed using practices drawn from established safety incident reporting programs1 vulnerability disclosure programs1 and programs to identify and develop lessons learned2 7#b7 emphasize non-punitive reporting1 protection of sensitive and proprietary information1 and dissemination of lessons learned1 as appropriate2 and 7#c7 include a mechanism to enable timely access to and sharing of relevant logs1 system data1 and model information as necessary to support analysis and response4 7d7 ,designation of official ,the ,secretary shall designate an appropriate official for the reporting1 tracking1 analysis1 and remediation of covered ,,ai incidents and covered ,,ai vulnerabilities under this section4 ,the ,secretary1 acting through such official1 shall receive and standardize reports1 conduct trend analysis1 identify recurring risks and failure modes1 and issue guidance1 alerts1 and recommendations1 as appropriate4 7e7 ,reporting and categorization 7#a7 ,the ,secretary shall require prompt reporting to the official designated under subsection 7d7 of-- 7,a7 any covered ,,ai incident2 and 7,b7 any covered ,,ai vulnerability4 7#b7 ,the ,secretary1 acting through the official1 shall categorize each incident or vulnerability reported to the official according to whether the incident or vulnerability requires-- 7,a7 a ,department-wide response2 7,b7 a response at the program level2 or 7,c7 a response at a local level4 7f7 ,department-,wide and program-,level matters 7#a7 ,in the case of any incident or vulnerability categorized under subsection 7e77#b77;,a7 or 7,b71 the ,secretary1 acting through the official designated under subsection 7d71 shall coordinate any responses that the ,secretary considers appropriate1 such as remediation1 retesting1 mitigation measures1 or deployment restrictions4 7#b7 ,in addition1 in the case of any incident or vulnerability described in subsection 7e77#b77;,a71 the ,secretary1 acting through the official1 shall require-- 7,a7 a documented corrective action plan2 and 7,b7 validation that the mitigation measures1 if any1 in such plan have been implemented before continued operational use4 7g7 ,protection of reports 7#a7 ,the ,secretary shall establish a protected disclosure process1 informed by established vulnerability disclosure practices1 through which members of the ,armed ,forces1 civilian employees1 contractors1 and subcontractors at any tier may report covered ,,ai incidents and covered ,,ai vulnerabilities in good faith4 7#b7 ,the ,secretary shall ensure that a person making a report in good faith under paragraph 7#a7 is not1 on the basis of that report alone1 subject to adverse contract action1 subject to adverse personnel action1 or otherwise retaliated against by the ,department4 7h7 ,protection of information ,the ,secretary shall establish procedures to protect sensitive1 proprietary1 and classified information submitted through the protected disclosure process under subsection 7g74 7i7 ,annual report 7#a7 ,in each of years #bjbg through #bjca1 the ,secretary shall submit to the congressional defense committees an annual report on the program4 ,the report shall include-- 7,a7 the number of reports made of incidents and vulnerabilities and the categorizations of such reports2 7,b7 a summary of significant trends1 recurring risks1 systemic issues1 and corrective actions taken in response2 7,c7 in the case of any covered ,,ai incident resulting in the loss of life of1 or in bodily harm to1 a member of the ,army1 ,navy1 ,marine ,corps1 ,air ,force1 or ,space ,force-- 7i7 a description of the incident1 including the system or systems involved and the operational context2 7ii7 the date and time the incident occurred2 7iii7 an assessment of the cause and operational consequence of the incident2 and 7iv7 any corrective actions taken2 and 7,d7 any recommendations for changes to testing1 procurement1 cybersecurity1 or deployment policies relating to artificial intelligence systems4 7#b7 ,each report under this subsection shall be submitted in unclassified form but may include a classified annex4 7j7 ,definitions ,in this section3 7#a7 ,the term artificial intelligence has the meaning given such term in section #ejjb of the ,national ,artificial ,intelligence ,initiative ,act of #bjbj 7#ae ,u4,s4,c4 #idja74 7#b7 ,the term covered ,,ai incident means an event in which an artificial intelligence system-- 7,a7 causes unintended operational1 safety1 or security harm2 7,b7 operates outside authorized parameters or approved safety1 legal1 or mission guardrails2 7,c7 materially degrades mission performance or reliability in a real-world or operationally representative environment2 7,d7 fails to respond to an operator disengage command2 7,e7 operates in a manner that1 under reasonably foreseeable circumstances1 could have resulted in significant unintended operational1 safety1 or security harm2 or 7,f7 operates in a manner that raises concerns regarding system control and autonomy4 7#c7 ,the term covered ,,ai vulnerability means an exploitable weakness1 vulnerability1 or systemic issue in an artificial intelligence system or related component that could materially affect mission performance1 compromise system integrity1 create safety risk1 or result in unauthorized or unintended behavior4 4