Artificial intelligence technology, public trust, and effective governance

Advancement in information technology continues to evolve especially the field of artificial intelligence (AI). Research studies have been conducted evaluate perceptions Americans on development and utilization AI if it is appropriate use public administrative duties. The research revealed that society fragmented regarding acceptance AI, whether decisions could long-term effects labor industry, legal system, national security. 2018 Public Opinion Survey significant concerns among American yet also a recognition its promise. goal this article further develop governance framework for considers importance trust policy. First, discusses necessity effective emergent technology. Then, evaluates opinion specifically pertains governance. concludes with discussion why central good 信息技术的进步不断发展，特别是在人工智能 (AI)领域。一系列研究已评价了美国人对AI技术开发和利用的看法，以及在公共行政职责中使用AI是否合适。研究表明，就AI的接受度，以及AI决策能否对劳动力产业、立法系统和国家安全产生长期影响而言，社会的看法是碎片化的。“2018人工智能民意调查”揭示了美国公众对AI的显著担忧，但也揭示了对AI承诺的认可。本文旨在进一步提出一项AI治理框架，将公众信任对AI政策的重要性考虑在内。第一，本文探讨了公众信任对于新兴技术的有效治理的必要性。第二，本文评价了公众对AI技术治理的看法。本文的结论探讨了为何公众信任对于AI善治至关重要。 El avance en la tecnología de información continúa evolucionando, especialmente el campo inteligencia (IA). Se han realizado estudios investigación para evaluar las percepciones los estadounidenses sobre desarrollo y utilización IA si es apropiado utilizar tareas administrativas públicas. La reveló que sociedad está fragmentada con respecto aceptación decisiones podrían tener efectos largo plazo industria laboral, sistema seguridad nacional. Encuesta Opinión Pública preocupaciones significativas entre público estadounidense IA, pero también un reconocimiento su promesa. objetivo este artículo seguir desarrollando marco gobernanza considere importancia confianza pública política IA. En primer lugar, analiza necesidad eficaz emergente. Luego, evalúa opinión se relaciona específicamente gobernanza. concluye una discusión por qué fundamental buena Answering these questions will help organizations better understand how their choices regulate may or hinder ability deliver shared values (Schiff et al., 2021). Furthermore, we advance an argument should be scholarly efforts A recent review existing frameworks rarely discussed, much less emphasized, by scholars (Robles & Mallinson, 2023, Under Review). However, vital embrace technologies (Zuiderwijk Thus, argue need pay more attention trust. Ours not first analysis public, but instead considering overall preferences desires versus human services (Gaozhao 2022; Zhang Dafoe, 2020), focus extent which trusts particularly when government involved. explicitly tie interest developing governing AI. In doing so, highlight gaps knowledge filled understanding can build sustain regulations private sector. This study begins literature regulation. It then as largely ignore concept followed specific has done opinion. Following review, paper uses representative sample attitudes including fairness hiring, criminal justice, data privacy. such inform model United States myriad additional are raised Emergent receive reception citizens involved policy (Flynn 2012; Flynn Bellaby, 2007). continued emergence new (e.g., hydrogen fuel, nanotechnology, more) presents interesting challenge where institutions reap potential benefits, do so they bolster (Macoubrie, 2006; Paulsen, case, sociocultural phenomenon having inclusion open conversation—including expressing risk misunderstandings–increases probability embraced Hydrogen energy, wind farms, carbon capture storage, nanotechnology each represent opportunities reduce dependency fossil fuels, improve environmental conditions, economy, and, some instances, foster conversation around risks hazards (Bäckstrand 2011; Dwyer Bidwell, 2019; Macoubrie, Sandaka Kumar, 2023). But faces concerns. example, produced using modern industrial methods safety protocols, surround flammability 2012). To make matters worse, does offer stable explanation mitigation (Sandaka Not mention past disasters, like explosion Hindenburg, even professional, imagination These issues misunderstandings led rejection Another generation electricity farms coastal areas high-altitude hills. At look, generating appears benevolent. citizen become critical topics concern. energy gap resulted from lack policy-making process (Dwyer 2019). engagement important element gaining preventing NIMBYism.1 organization leaders take initiative generate dialogues community leverage network solve complex problems lesson learned makers value procedural during planning sources through (Liu 2020). Alas, approach often used siting turbines, resulting local controversy (Olsen, 2010). Carbon storage another example Kim 2016). Climate activists remain at odds over fuels must eliminated source climate change without dramatic shift Kalkuhl There remains insufficient clarity across subnational governments uncertainties economic impacts (Kalkuhl As result socio-technical socio-political issues, balanced 2011). is, however, evidence pushed manipulated lobbies motivated profit 2020; Nanotechnology stands revolutionize many products, water filtration systems, biotechnology, computer technology, fabrics, other applications (Boholm Larsson, Todaro Like there relatively little nano-research implications 2006). present collaboration dispositions nanotechnology's notable similar related (Todaro applications, scientists different fields legal, management, communication) investigating Experts innovative consider citizens' thus, continue communication technologies, well creates what researchers “blind spots” processes (Veen 2010, p. 810). “Blind refer leads overlooked For was patients celiac disease (CD) observe individual reactions medical treatment scientific excited technological breakthrough. mixed patient sentiments desire participate treatment. Researchers discovered several might mistrust means presentation number one factor contributing mistrust, limited Likewise, fail warm welcome due clear engagement. Such disregard (Charalampia 2022). Granted, uphill battle. formation before (Druckman Bolsen, true visions futures already shaped television movies (Brewer Murdoch 2011), sentient (Dieter Gessler, Individuals personal heuristics, level science cost, toward tend health possible life hazardous accidents Broadly speaking disconnect between Huang, mindful changes easily influenced (Huang, success any policy, requires building sectors depend inclusion, dialogue, transparent sharing (Wilson, pointed out, opposite secrecy commonly shroud advances sector (Pasquale, Regardless motivations, managers realize positive attitude key agenda move forward (Gutierrez Gaviria, Wilson, bureaucrats shape contours service delivery citizens. They directly assist making actions Before introduction into services, generally provided efficient few performance were always positive, traditional bureaucracy viewed slow lagging. advancement machine learning (ML) algorithms assisting servants faster providing (Amann 2023; Gaozhao Moreover, spends time electronic than ever Meijer Boon, Without knowledge, replacing servant front-line e-government about (Ansell Miura, amount collected emails, system logs, search queries websites, records quickly effectively while helping (Ahmed Kemper Kolkman, transition replacement changed respect institution quality Quinn apply social benefits healthcare services), guidance robotic voice systems mobile powered (Quinn Online welfare being processing verifying approving rejecting (Eubanks, 2018; cases, rely computerized decisions. made biased affecting individuals' benefits. occurrences cause emotional stress unnecessary frustration, both thus crossroads fast, efficient, accessible maintaining (Kemper evident previous heavy reliance resources needs updated cope growing demands adapt Scholars considered symbolic representation passive accomplished based algorithmic accountability, transparency, fair practices bias remains, limitations proposed far fallen short recognizing systematic wanting Most rooted 1948 Universal Declaration Human Rights, adopted after World War II (Donahoe Megan MacDuffee, follow recently developed principles guidelines Nation's Guiding Principles Business Rights include privacy protection, well-being, fairness, (Agbese echo multiple internationally, nationally, sectors. only implies principle (Can, because coherent program clarion call (Taeihagh, noted above, most elements noticeable received enough case One spot” democratic governance, positively negatively affect institutions, those intentions demand reliable, safe used, harm (Ramlaoui Semma, 2014). Additionally, shown promise wicked problems, combating COVID-19 pandemic (Valle-Cruz Sandoval-Almazán, Technological advancements, history negative consequences, pays price (Krarup Horst, lead dissatisfaction, loss authorities, standing (Standford, principle. know views technology? tends ignored implementation Applications (Criado Gil-Garcia, Kieslich 2021), advancements come consequences. general people left behind progresses (Kieslich sentiments, represented (Zhang Governments “people happily accept though care nor all” 2021, 1563). believes widespread driven companies expansion believe sufficient supporting organizations' adoption services. Oftentimes, consumers purchase tech industry diverse persuasive methods, target advertising, influence emotions adopting technologies. immune pressures fact, under increasing pressure same time, message challenges delivering goods resolved Circumstances people's (Wenzelburger Citizens pleased flexibility access devices, online automated (Bucher, Wenzelburger known behave erratically, discriminatory decisions, require adjustments reveals concerns, efficiency (Shin-Cheng effectively; people, minoritized communities, experience racial discrimination perceive ethical While generally, now turn evaluating assess nationally survey Artificial Intelligence: Attitude Trends 2020).2 secondary dataset publicly available researchers. includes 2000 responses, societal implications, demographics, gender, age, race, marital status, ideology, political party, education. We derive practical common method (Kock Procedural controls original our control bias. independent dependent variables (a bias), design gave respondents instructions, ensured anonymity, avoided (Podsakoff Because secondarily, causal linkages would 2021) proximal separation actual Harman's single test all measures found accounted 14% variance, below 50% concern response (Harman, 1960; Kock 2021; Podsakoff 2003). sought multifaceted Each chosen links begin assessing feelings safe, trustworthy, fair, transparent. examine justice hiring. presumption automate unfair. transparency hiring court measured questions. automation transmit information, Personal Identifiable Information (PII), digital format. PII individual's driver license, security, birth certificate, residential address, sensitive history. Next, support development. civil rights. finally broaden possibility causing harmful humans, comprehend protects society. opinions differ party affiliation, education, race.3 nominal ordinal categories. Table 1 descriptive statistics variables. two flavors. Either asked confidence likely variety arise. substantial variation protection lowest (i.e., least likely) average response, whereas rights highest likely). Across variables, privacy, consequences created (Wirtz series logistic regressions estimated POLR R analyses. examining vary demographic groups, useful Figure displays distributions eight find plurality expresses “a confidence” majority no (52% combined). More (46% 52%, respectively) unlikely (30% 23%, respectively). full 68% expect Perhaps unsurprisingly, 74% view “very important.” makers, consider, person supports (46%), 30% ambivalent. (59%) Finally, AI's pessimistic. bright spots say low especially, relevant Turning assessment population, 2 statistical separate models, variable. table reports results ratios. When ratio greater 1, variable increases moving up scale. decrease. Alternatively, increased down keep mind scales variable, which, reference, depicted 1. Higher levels education rights, placing respondent moves scale 14%, 12%, 16%. differences white non-white Non-white 46% viewing important, 79% higher likely. Female 32% males safe/trustworthy, 37% development, 35% unlikely. correlation opinion, ideology does. Increasing conservatism relates safe/trustworthy (29% increase odds), likelihood (13% opposition (17% odds). aim emerging end, extensive covering range 2023), narrowed Our highlights various organizations, within substantiv