,h4 ,,res4 #abej ,in the ,house of ,representatives1 ,u4 ,s41 ,december #h1 #bjbj ,,resolution #a4 ,guiding principles of the national artificial intelligence strategy of the ,united ,states 7a7 ,findings ,the ,house of ,representatives finds the following3 7#a7 ,in general1 artificial intelligence is the ability of a computer system to solve problems and to perform tasks that would otherwise require human intelligence4 7#b7 ,artificial intelligence will transform the nature of work and nearly all aspects of the ,united ,states economy4 7#c7 ,artificial intelligence will have immense implications for the security of the ,united ,states and its allies and partners4 7#d7 ,investments made by the ,united ,states ,government will be instrumental in the research and development of artificial intelligence and artificial intelligence-enabling technologies1 as it has been for many of the world's revolutionary technologies4 7#e7 ,developing and using artificial intelligence in ways that are ethical1 reduce bias1 promote fairness1 and protect privacy is essential for fostering a positive effect on society consistent with core ,united ,states values4 7#f7 ,the ,obama ,administration released the ,big ,data ,research and ,development ,initiative in #bjab1 ,executive ,order #acgjb 7relating to creating a national strategic computing initiative7 in #bjae1 and the ,national ,artificial ,intelligence ,research and ,development ,strategic ,plan in #bjaf4 7#g7 ,the ,trump ,administration released ,executive ,order #achei 7relating to maintaining ,american leadership in artificial intelligence71 updated the ,national ,artificial ,intelligence ,research and ,development ,strategic ,plan in #bjai1 and released ,office of ,management and ,budget guidance for regulation of artificial intelligence applications in #bjbj4 7#h7 ,in ,may #bjai1 the ,organisation for ,economic ,co-operation and ,development 7,,oecd7 adopted the ,,oecd ,principles on ,artificial ,intelligence1 which included the principles of inclusive growth1 sustainable development and well-being1 human-centered values and fairness1 transparency and explainability1 robustness1 security and safety1 and accountability4 7#i7 ,in ,february #bjbj1 the ,european ,commission began a consultation process with the release of their white paper ,on ,artificial ,intelligence - ,a ,european approach to excellence and trust1 which set out policy options for a coordinated ,european approach to artificial intelligence regulation4 7#aj7 ,in ,june #bjbj1 the ,g#g and several partners launched the ,global ,partnership on ,artificial ,intelligence to increase cooperation focused around the areas of responsible artificial intelligence1 data governance1 the future of work1 and innovation and commercialization4 7#aa7 ,several ,united ,states allies1 including ,canada1 ,denmark1 ,estonia1 ,france1 ,finland1 ,germany1 the ,netherlands1 and ,south ,korea1 have published national artificial intelligence strategies with detailed funding commitments4 7#ab7 ,in #bjag1 ,china published a national artificial intelligence strategy that detailed the ,chinese ,communist ,party's goal to become the world's primary artificial intelligence innovation center by #bjcj4 7#ac7 ,in #bjai1 ,russia published a national artificial intelligence strategy and1 in #bjag1 ,russian ,president ,vladimir ,putin said that whoever becomes the leader in this sphere will become the ruler of the world4 7#ad7 ,in #bjah1 the ,subcommittee on ,information ,technology of the ,committee on ,oversight and ,government ,reform of the ,house of ,representatives1 under the leadership of ,chairman ,will ,hurd and ,ranking ,member ,robin ,kelly1 published ,rise of the ,machines3 ,artificial ,intelligence and its ,growing ,impact on ,u4,s4 ,policy following a series of hearings on artificial intelligence with experts from academia1 industry1 and government1 concluding that the ,united ,states cannot maintain its global leadership in artificial intelligence absent political leadership from ,congress and the ,executive ,branch4 7#ae7 ,congress serves a critical role in establishing national priorities1 funding scientific research and development1 supporting emerging technologies1 and sustaining cooperation with our allies to protect the national security of the ,united ,states4 7b7 ,national artificial intelligence strategy principles ,it is the sense of the ,house of ,representatives that the following principles should guide the national artificial intelligence strategy of the ,united ,states3 7#a7 ,global leadership4 7#b7 ,a prepared workforce4 7#c7 ,national security4 7#d7 ,effective research and development4 7#e7 ,ethics1 reduced bias1 fairness1 and privacy4 #b4 ,global leadership ,it is the sense of the ,house of ,representatives that the ,united ,states should take a global leadership role in artificial intelligence4 #c4 ,workforce preparation 7a7 ,findings ,the ,house of ,representatives finds the following3 7#a7 ,artificial intelligence and automation will present significant challenges to workers in affected industries due to the automating of some routine and repetitive tasks1 but will also create additional employment opportunities4 7#b7 ,closing the artificial intelligence talent gap in the short and medium-term will require a targeted approach to identifying and filling roles that require the skills to build and work with artificial intelligence systems4 7#c7 ,the ,united ,states should take a leadership role in the artificial intelligence-driven economy by filling the artificial intelligence talent gap and preparing ,united ,states workers for the jobs of the future1 including by prioritizing inclusivity and equal opportunity4 7#d7 ,departments and agencies of the ,federal ,government are increasingly using data to administer benefits1 assess outcomes1 and fulfill other mission-critical activities4 7#e7 ,effectively creating1 managing1 and implementing artificial intelligence related research and development grants will require technical expertise4 7#f7 ,departments and agencies of the ,federal ,government will need to be able to recruit employees with technical expertise4 7#g7 ,lifelong learning and skill acquisition can increase flexibility with respect to career opportunities4 7#h7 ,the ,united ,states will need to be able to attract the best artificial intelligence researchers and computer scientists from around the world to work in the ,united ,states4 7b7 ,matters to consider 7#a7 ,education ,it is the sense of the ,house of ,representatives that the national competitiveness of the ,united ,states in artificial intelligence would benefit from-- 7,a7 increased funding for ,federal programs that support science1 technology1 engineering1 mathematics1 and computer science education2 7,b7 grant programs that continue funding the integration of ethics courses and modules into science1 engineering1 and computer science curricula2 7,c7 new education programs of study related to artificial intelligence that incorporate industry-recognized credentials1 including certifications and certificates1 embedded within secondary and postsecondary degree programs2 and 7,d7 continued support for teacher preparation programs that increase the number of teachers with the ability to teach science1 technology1 engineering1 mathematics1 and computer science education4 7#b7 ,promoting diversity ,it is the sense of the ,house of ,representatives that-- 7,a7 the inclusion of students from historically under-represented groups in existing technology education programs would benefit a diverse artificial intelligence workforce2 and 7,b7 recruitment and retention policies with respect to under-represented communities and marginalized groups in the ,federal workforce should be reviewed for the purpose of determining if such policies require modification for technology workers4 7#c7 ,artificial intelligence training 7,a7 ,in general ,it is the sense of the ,house of ,representatives that the ,federal ,government should assess the effectiveness of current public workforce development programs with respect to the additional support such programs will need to effectively address job disruptions and job creations that result from the increased use of artificial intelligence4 7,b7 ,work-based learning and on-the-job training programs ,it is the sense of the ,house of ,representatives that the ,federal ,government should support the adoption of work-based learning and on-the-job training programs to prepare the ,united ,states workforce for an artificial intelligence-influenced economy1 including by-- 7i7 undertaking studies to determine best practices to implement such programs2 and 7ii7 ensuring that there is sufficient ,federal funding to support high-quality programs that coordinate with ,federal workforce development programs4 7#d7 ,federal hiring practices ,it is the sense of the ,house of ,representatives that the ,federal ,government should-- 7,a7 allow technical experts to use their skills to assist multiple departments and agencies of the ,federal ,government1 such as the ,united ,states ,digital ,service2 7,b7 focus on the retention of non-partisan experts within the ,federal ,government who are working to modernize ,federal information technology2 7,c7 include in the criteria for recruiting for artificial intelligence jobs the consideration of a multi-disciplinary set of skills1 including an understanding of ethical practices with respect to the design and use of artificial intelligence systems1 privacy1 information security1 law1 and civil liberties2 7,d7 review hiring practices for employment in the ,federal ,government for the purpose of ensuring that such practices do not disqualify individuals with a less traditional background1 including due to a lack of undergraduate or graduate degree attainment1 who have skills that will benefit work in artificial intelligence systems management and research and development2 and 7,e7 conduct studies with respect to best practices for skills-based hiring4 #d4 ,national security 7a7 ,findings ,the ,house of ,representatives finds the following3 7#a7 ,artificial intelligence will have immense implications for national and international security4 7#b7 ,artificial intelligence tools and systems can augment human intelligence through human-machine collaboration and teaming across the national security ecosystem4 7#c7 ,ensuring that the public trusts the ability of the military to ethically use artificial intelligence and that human operators in human-machine teams trust the artificial intelligence will be critical factors with respect to the successful implementation of artificial intelligence systems4 7#d7 ,the continued proliferation of national artificial intelligence strategies1 plans1 statements1 and investments demonstrates the increase in global competition in this area4 7#e7 ,new paradigms will be required to effectively test artificial intelligence and to ensure that it is reliable and stable4 7#f7 ,export and investment controls will be important policy tools to prevent the acquisition of sensitive artificial intelligence and artificial intelligence-enabling technologies1 including hardware such as semiconductors and semiconductor manufacturing equipment1 by ,china1 ,russia1 and other adversaries4 7b7 ,matters to consider 7#a7 ,collaboration with foreign nations ,it is the sense of the ,house of ,representatives that the ,united ,states should-- 7,a7 leverage its alliances to promote democratic principles1 foster research collaboration1 and develop common standards with respect to artificial intelligence2 7,b7 promote the interoperability of artificial intelligence for the purpose of strengthening alliances2 7,c7 along with allies1 take a leading role in international forums to set artificial intelligence principles1 norms1 and standards2 and 7,d7 undertake efforts to engage with ,china and ,russia with respect to-- 7i7 shared concerns about artificial intelligence safety2 and 7ii7 confidence-building by establishing crisis communications procedures designed to reduce the likelihood of unintentional use and the risk of escalation with respect to artificial intelligence systems4 7#b7 ,foreign artificial intelligence capability ,it is the sense of the ,house of ,representatives that national security agencies should consider conditions-based and capabilities-based approaches when evaluating global artificial intelligence capabilities4 7#c7 ,development and deployment ,it is the sense of the ,house of ,representatives that national security agencies should-- 7,a7 collaborate with experts in academia1 the private sector1 and other departments and agencies of the ,federal ,government to develop best practices for testing1 evaluation1 validation1 and verification of artificial intelligence systems2 7,b7 devote agency resources1 including investing in research1 for the purpose of promoting trustworthiness with respect to human-machine teams2 7,c7 engage with experts to develop guidelines for the ethical development and use of artificial intelligence systems2 and 7,d7 prioritize the development of artificial intelligence systems to cover non-critical tasks until such systems can achieve suitable standards of reliability1 interoperability1 and security4 7#d7 ,export and investment controls ,it is the sense of the ,house of ,representatives that the ,united ,states should collaborate with its allies to prevent the misuse of artificial intelligence systems by ,china1 ,russia1 and other adversaries4 #e4 ,research and development 7a7 ,findings ,the ,house of ,representatives finds the following3 7#a7 ,federal funding plays an important role in research and development4 7#b7 ,federal research and development investments need to be significantly increased to ensure ,united ,states leadership in artificial intelligence4 7#c7 ,federally supported research will play an important role in supporting artificial intelligence techniques that are critical to ,united ,states artificial intelligence leadership1 including by exploring novel techniques that leverage smaller data sets to train artificial intelligence systems and making more efficient use of computing resources4 7#d7 ,artificial intelligence advances are enabled by ,federal research and development investments in other technology sectors because ,united ,states economic competitiveness and national security will depend on strong capabilities across a range of technologies4 7#e7 ,computing power is essential to progress in artificial intelligence development1 and the amount of computing power required for artificial intelligence training runs is increasing exponentially4 7#f7 ,a new wave of technological advances could be fostered by combining and increasing access to government-owned and government-funded computing and data resources4 7#g7 ,expanding access to digital infrastructure1 such as broadband1 will be essential to creating new job opportunities and stimulating the growth of new technology and innovation clusters to support ,united ,states leadership in artificial intelligence4 7#h7 ,incentivizing research and development across the private sector1 particularly from smaller companies1 will further strengthen the ,united ,states innovation ecosystem4 7#i7 ,the ,united ,states is an attractive research and development partner because it is home to world-class universities1 research institutes1 and corporations4 7#aj7 ,decades of experience show that joint work with foreign researchers can be done with great benefit and little detriment to ,united ,states economic and national security with the implementation of proper safeguards4 7#aa7 ,artificial intelligence standards and measurement are essential to fostering artificial intelligence technologies that are safe1 secure1 reliable1 and comport with the norms and values of the ,united ,states4 7#ab7 ,metrics are how the artificial intelligence research community guides itself and prioritizes research4 7#ac7 ,benchmark tests are necessary to understand the performance of an artificial intelligence system4 7#ad7 ,current tests for measuring artificial intelligence range from vague and conceptual to well-defined and mature4 7#ae7 ,artificial intelligence measurement methodologies are not static and will require periodic reexaminations and updates of testing methodologies to ensure that artificial intelligence systems are functioning according to best-known practices4 7#af7 ,united ,states leadership in global artificial intelligence standards-setting will help ensure that artificial intelligence implementations are in accordance with ,united ,states strengths and comport with the interests and values of the ,united ,states4 7#ag7 ,public engagement is necessary for developing voluntary consensus standards1 guidelines1 and frameworks to ensure diverse perspectives are considered4 7b7 ,matters to consider 7#a7 ,federal funding ,it is the sense of the ,house of ,representatives that the ,federal ,government should increase investments in artificial intelligence research and development and related fields4 7#b7 ,collaboration with other entities ,it is the sense of the ,house of ,representatives that departments and agencies of the ,federal ,government should collaborate-- 7,a7 with the private sector1 civil society1 and academia-- 7i7 to ensure that the ,united ,states innovation ecosystem leads the world in artificial intelligence research and development2 and 7ii7 to develop voluntary consensus standards1 guidelines1 and frameworks that will help create shared conceptual foundations1 terminology1 and best practices for artificial intelligence fairness and bias mitigation2 and 7,b7 with science funding organizations in like-minded countries to establish multilateral teams of artificial intelligence researchers from the public and private sectors to promote additional talent development and foster partnerships on artificial intelligence research and development4 7#c7 ,expanding digital access ,it is the sense of the ,house of ,representatives that the ,federal ,government should-- 7,a7 expand access to broadband in rural and underserved areas2 7,b7 expand the availability of affordable graphics processing units and high-performance computers in rural and underserved areas2 7,c7 improve digital infrastructure in the ,united ,states2 and 7,d7 make data created by federally-funded scientific and technical research publicly available with appropriate privacy protections to provide artificial intelligence researchers with new data sets to train their systems4 7#d7 ,national computing and data resource ,it is the sense of the ,house of ,representatives that ,congress should consider establishing a national computing and data resource4 7#e7 ,access to ,national ,laboratories ,it is the sense of the ,house of ,representatives that the existing supercomputing labs at the national laboratories and technology centers of the ,department of ,energy should expand opportunities for academics and researchers to access such labs for artificial intelligence research and research related to artificial intelligence4 7#f7 ,tax incentives ,it is the sense of the ,house of ,representatives that ,congress should examine whether targeted incentives and reforms to the ,internal ,revenue ,code of #aihf would increase private sector research and development1 particularly with respect to small cap corporations4 #f4 ,ethics1 reduced bias1 fairness1 and privacy 7a7 ,findings ,the ,house of ,representatives finds the following3 7#a7 ,the rise of artificial intelligence has great potential to improve quality of life for individuals in the ,united ,states1 provided it is developed and used in a manner that is ethical1 reduces bias1 promotes fairness1 and protects privacy4 7#b7 ,a diverse artificial intelligence workforce is important for mitigating bias4 7#c7 ,the ,united ,states is uniquely positioned to leverage its diverse workforce to lead in artificial intelligence4 7#d7 ,the starting point for ,federal oversight of artificial intelligence should be to review existing regulatory frameworks4 7#e7 ,regulatory sandboxes1 in general1 refer to regulatory structures where a participant obtains limited or temporary access to a market in exchange for reduced regulatory uncertainty1 and can be used to test a product designed to mitigate unintended bias or promote fairness in a small-scale environment and under the supervision of regulators4 7#f7 ,federal programs should have necessary safeguards and oversight processes4 7#g7 ,artificial intelligence regulatory approaches should consider the level of risk associated with different artificial intelligence applications4 7b7 ,matters to consider 7#a7 ,bias mitigation ,it is the sense of the ,house of ,representatives that departments and agencies of the ,federal ,government should-- 7,a7 support technical and non-technical research and development to address potential bias1 fairness1 and privacy issues in artificial intelligence2 7,b7 improve access to a broad range of non-sensitive government data assets to help train artificial intelligence systems2 7,c7 implement title ,,ii of the ,foundations for ,evidence-,based ,policymaking ,act of #bjah 7,public ,law #aae-dce2 #acb ,stat4 #eebi72 7,d7 develop policies to identify the data used to train artificial intelligence algorithms as well as data analyzed by artificial intelligence algorithms and systems in use by departments and agencies2 and 7,e7 further develop and release to the public available benchmark data assets with the proper safeguards to protect privacy1 mitigate bias1 and promote inclusivity4 7#b7 ,regulation and legislation review ,it is the sense of the ,house of ,representatives that congressional committees should-- 7,a7 review the range of existing ,federal regulations and laws that potentially apply to artificial intelligence2 7,b7 determine which laws apply to artificial intelligence2 7,c7 determine if any gaps in appropriate legislation and regulation exist and how such gaps could be addressed2 7,d7 advance ,federal privacy reforms that build trust1 prevent harm1 and maintain ,united ,states global leadership in artificial intelligence2 and 7,e7 conduct regular oversight of artificial intelligence policies in the executive branch within their jurisdiction4 7#c7 ,federal ,funding ,it is the sense of the ,house of ,representatives that ,congress should support funding for departments and agencies of the ,federal ,government interested in adopting programs1 including regulatory sandboxes1 for the purposes of testing artificial intelligence tools in limited markets4 ,clerk4