First Monday

Chief information officers' perceptions about artificial intelligence: A comparative study of implications and challenges for the public sector by J. Ignacio Criado, Rodrigo Sandoval-Almazan, David Valle-Cruz, and Edgar A. Ruvalcaba-Gomez



Abstract
This article presents a study about artificial intelligence (AI) policy based on the perceptions, expectations, and challenges/opportunities given by chief information officers (CIOs). In general, publications about AI in the public sector relies on experiences, cases, ideas, and results from the private sector. Our study stands out from the need of defining a distinctive approach to AI in the public sector, gathering primary (and comparative) data from different countries, and assessing the key role of CIOs to frame federal/national AI policies and strategies. This article reports three research questions, including three dimensions of analysis: (1) perceptions regarding to the concept of AI in the public sector; (2) expectations about the development of AI in the public sector; and, (3) challenges and opportunities of AI in the public sector. This exploratory study presents the results of a survey administered to federal/national ministerial government CIOs in ministries of Mexico and Spain. Our descriptive statistical (and exploratory) analysis provides an overall approach to our dimensions, exploratory answering the research questions of the study. Our data supports the existence of different governance models and policy priorities in different countries. Also, these results might inform research in this same area and will help senior officials to assess the national AI policies actually in process of design and implementation in different national/federal, regional/state, and local/municipal contexts.

Contents

Introduction
Research background: Artificial intelligence in the public sector
Analytical strategy and research methods
Results
Discussion
Conclusion

 


 

Introduction

The use of artificial intelligence (AI) in government is not a novelty. Slone (1991) reported a first approach in the U.S. Navy two decades ago. Nowadays, more powerful computer processors, and the fast evolution of hardware and software industries have fostered an important development of AI in different sectors (Brynjolfsson and McAfee, 2017; Bucher, 2017; Russell and Norvig, 2016). Governments and public administrations have adopted new practices of AI in different areas, including personalized public services in health, education, security, or defense (Agarwal, 2018; Desouza, 2018; Janssen and Kuk, 2016; Pencheva, et al., 2020; Sun and Medaglia, 2019). The Netherlands established an expert group to study options on AI in defense (De Spiegeleire, et al., 2017). Probably, the most common use of AI in government occurs in security, using facial recognition. Gershgorn (2019) reported facial recognition by California police; Margetts and Dorbantu (2019) described the case of London Metropolitan Police in 2017. National governments around the world have defined AI strategies and policies facing an uncertain future (Araya, 2019; Dutton, 2018; Lee, 2018). Therefore, the aim of this paper is, like in other domains of Internet based technologies in government (Ahmad, et al., 2019; Janssen, et al., 2019; Sideri, et al., 2019), understanding the perceptions, expectations, challenges/opportunities of the main public sector stakeholders: CIOs (in two different national governments) about AI in government. CIOs are decision-makers directly involved in ministries implementing applications, cases, and techniques based on AI. Their ideas and visions are critical during this initial stage to shape and guide the future of AI in the public sector.

Despite the growing importance of AI in the public sector, there are few studies addressing this emergent topic. Valle-Cruz, et al. (2020, 2019) expanded a literature review on AI in government, noting that adoption and implementation practices advance faster than scientific and theoretical reflection. In general, publications about AI in the public sector relies on experiences, cases, ideas, and results from private sector companies (Barth and Arnold, 1999; Seaver, 2013; Zhang and Dafoe, 2019; Zuboff, 2015). From the perspective of public administration, some argue that the intensive utilization of social media, robotics, big data, and lately, AI in the public sector could be addressed as the fourth wave of an information and communication technologies (ICTs) evolution in the public sector or the implementation of the fourth industrial revolution in government (Criado and Gil-Garcia, 2019; Criado, et al., 2020; Valle-Cruz, 2019; Criado, et al., 2013; Mergel, et al., 2016; Meijer, 2018). This new wave of technological diffusion in the public sector includes all functional dimensions (i.e., strategic management, human resource management, performance assessment, institutional communication) (Barth and Arnold, 1999), and policy areas (i.e., health, education, control of borders, customer service, emergencies, tax, social benefits) (Wirtz, et al., 2019), and is based on the volume of open and big data and the new processing capacities of organizations (Burrell, 2016; Cukier and Mayer-Schoenberger, 2013; Margetts, 2017). Hence, this reality opens the door in public sector organizations to innovate in different dimensions.

Traditionally, CIOs have been a source of knowledge about the opportunities, challenges, and problems of implementing emerging technologies and trends in government (Ganapati and Reddick, 2012). Methodologically, this article is based on a survey administered to national/federal CIOs in two countries (Mexico and Spain). These two countries are examples of cases in North American and European emergent systems of AI governance. National/federal governments are actors in different contexts to introduce new regulations about key issues, including protection and security of data, privacy of citizens, future of public employment, robotics implementation, automation of decision-making, ethics of algorithms, among others (Lee, 2018; Wright, 2018). We assume that not all countries are regulating and designing AI in government with the same principles, democratic values, and goals, above all, in China, European Union countries, or the United States (including its North American neighbors).

Therefore, the purpose of this study is to understand CIOs’ perceptions, expectations, and challenges/opportunities about AI in the public sector. This article presents three research questions, including three dimensions of analysis: (1) What is the perception of those who manage AI about this concept? Perceptions regarding to the concept of AI in the public sector; (2) What are the expectations of those who manage AI regarding its development in public administration? Expectations about the development of AI in the public sector; and (3) What are the main challenges and opportunities for public administrations managing AI? Challenges and opportunities of AI in the public sector. Our descriptive statistical analysis will provide an overall approach to our dimensions of study, exploratory answering our research questions.

The remainder of the article is as follows. The next section presents our research background, including our approach to the literature about AI in the public sector, a review of the emerging governance models of AI in the public sector (in Mexico and Spain), and the role of public managers’ perceptions about AI in government. The third section advances analytical strategy and research methods of the study, including our research questions. The fourth section presents the results of the study regarding our three dimensions of analysis. In the fifth section, we discuss our findings looking at the emergent literature of AI in government. The final section completes the article with the conclusion, developing ideas for the future development of this emergent area of research.

 

++++++++++

Research background: Artificial intelligence in the public sector

Artificial intelligence (AI) and algorithms are changing the way we think about the opportunities, challenges, and future of government and public sector organizations. Some authors suggest AI will transform the interaction with citizens and other societal actors, using algorithms and other technological instruments to augment these experiences from traditional services based on Web sites (Bucher, 2017; Desouza, 2018; Giest, 2017). Other scholars say AI will provide efficiency gains in public organizations derived from the generalization of huge sources of data and analytics to optimize all internal processes and activities (Williamson, 2014). Others assume that AI will inaugurate a different stage of the governance of public institutions including the disruptive nature of technologies related with AI in policy-making and decision processes (Pencheva, et al., 2020). Most public administration scholars assume that dynamics and ends of private companies can be immediately transferred to the public sector. In this section, our approach to AI and the public sector assumes the need to define a singular focus to employees, functions, citizens, and ends participating in public management and policy. From this starting point, we introduce the situation of AI in two different countries (Mexico and Spain) and, then, we highlight why it is important studying perceptions and expectations of senior public managers (CIOs) shaping the adoption and implementation of AI in government.

Looking for an approach to AI in the public sector

The literature about AI and algorithms is relatively limited specifically related to public administration. In particular, automation technologies are expected to have a direct impact in the case of the public sector. This would mean the transition from the interest in the automation of tasks, computerization policies, and digital governance (based on Web-mediated technologies), to a smart governance that require continuous interaction of and learning with humans (based on algorithm-mediated technologies) (Brynjolfsson and Mitchell, 2017; Margetts, 2017). Also, this new technological wave in the public sector would be characterized by the emergence of transformative organizational types (i.e., holacracy), or open cultures of collaboration among people and public employees (Criado and Gil-Garcia, 2019; Gil-Garcia, et al., 2018; Meijer, et al., 2019).

Despite previous ideas, our research focuses on the analysis of AI from non-deterministic perspectives of ICTs in government and public administration. AI in government involves the design, building, use, and evaluation of algorithms and computational techniques to improve the management of public agencies (Desouza, 2018). At this stage, governments all over the world are beginning to implement autonomous systems and techniques based on algorithmic governance to transform decision-making and operational processes, public service delivery, and interaction with citizens (Lazer, 2016; Margetts and Dorobantu, 2019; Meijer, 2018). Nonetheless, we do not have unanimous opinions about the implications of AI in the public sector.

Data nurturing AI are generated, collected, stored, and processed using information systems and algorithms that are technological devices seen as neutral, or at least more neutral that humans (Agarwal, 2018; Meijer, et al., 2016). These hyper-rational technological instruments are the foundation of descriptive, predictive, prescriptive, and automated positivist analytics (Vydra and Klievink, 2019). Thus, assumptions about AI are embedded in the logic of big data analytics, and the inscrutability of algorithms (Introna, 2015). These assumptions make it difficult for government organizations and public sector managers analyzing the risks of policy-making processes and public decisions based on biased sets of data or unethical algorithms (Ananny, 2016).

Moreover, the transformational effects of AI and algorithms on governments and public administrations remain controversial. Different AI models are under play and diverse policy entrepreneurs are collaborating to advance in the implementation of AI-based technologies. Among them, public administration CIOs are critical in understanding on-going policies, strategies, and initiatives. Therefore, understanding ideas and perceptions about AI and algorithms of those leading the adoption in the public sector is of paramount importance. However, we do not have comparative research studying different administrative contexts and approaches to AI.

Emerging governance models in the public sector: AI policies in Mexico and Spain

Artificial intelligence’s strategies and policies are developing quickly around the world. The Chinese government launched the most comprehensive and integrate strategy in 2017, including goals for education, skills acquisition, regulation, norms, among others (Dutton, 2018). In fact, the main intention was to become the world AI leader by 2030, with millions for acquisitions and industry mergers (Araya, 2019). On its side, the government of the U.S. has published in February 2019 a directive that included: innovation, industry, workers, and values but no financial investment for this effort. The European Commission (2018) proposed three stages to foster AI in public and private sectors (Dutton, 2018). From an analysis of official documents and policies, we found the idea of the emergence of different (Chinese, European and North American) models of AI governance (Wright, 2018). This conclusion was not only based on national priorities, strategies and policies, but also on different political, institutional, and public administration traditions.

Mexico and Spain have recently released national AI strategies, including actions plans oriented to government agencies and public administration. Mexico is in the middle of a historical change of government. The digital government scenario in the public sector in Mexico implies the consolidation of a comprehensive AI agenda. In June 2018, the report “Towards an artificial intelligence policy in Mexico: Taking advantage of the AI revolution” was commissioned by the British Embassy in Mexico and developed by C Minds and Oxford Insights in collaboration with the government of Mexico, in order to provide a first approach to this issue. On the other hand, a multisector team composed by practitioners, academic institutions, companies, startups, public agencies, and other key actors of the digital ecosystem and AI in Mexico called “Coalición IA2030Mx” launched a National Consultation of AI. This action had the objective of identifying the main challenges in Mexico regarding the digital transformation, specifically in AI. The survey was conducted from 15 August to 15 September 2018; 80 percent think that AI will have a positive effect on Mexican citizens. However, 45 percent believe that it will have a negative effect on data privacy and work transformation and so-called human replacement.

The case of Mexico exhibits efforts developed by academics, organizations of the civil society, and private companies to inaugurate an implementation of technological solutions based on AI in government. Particularly, this case is oriented to promoting leadership from private sector organizations, IT companies, and the massive utilization of personal data with minimum legal restrictions. These solutions include designs from the private sector, with implications for public employees and citizens on the development of private sector databases. This approach appears in line with other two North American countries, fueled by the NAFTA (Clarkson, 2008; Sapp, 2005), as well as in the field of digital government (Luna-Reyes, et al., 2010a, 2010b).

In the case of Spain, AI has monopolized the attention of scholars and private/public managers working on digitalization recently (European Commission, 2019). In particular, Spain follows the European Union approach on AI for algorithm-mediated governance, based on the protection of human rights, personal privacy, and personal data security (European Commission, 2018). This perspective has been stated in a recent national R+D+I strategy in AI (Spain. Ministry of Science and Universities, 2019), including the recommendations over AI with public administration data. Spain represents one case of the European way to imagine the future of AI and algorithms in government. Spain is in line with developments in AI in the European Union, as stated by the Joint Research Center of the European Commission (2020). That research indicated, from a sample of AI cases in the public sector, that “Portugal and The Netherlands hold 8% of total case studies each, followed at short distance by Denmark with 7% and then Estonia with 6% and France and Spain, both with 5% of the collected initiatives represented” (European Commission, Joint Research Center, 2020).

Open and public data needs to be modelled, analyzed, used, and exploited in a safe an ethical way by public organizations. This is intended to obtain the maximum benefit from open data using AI techniques, including efficiency and efficiency gains of their processes, and promoting public-private collaboration platforms to reduce costs. This model does not close the doors to private companies that may use open and public data with the aim of offering better public services, assuring social welfare, or public value. Therefore, the commitment of the Spanish government is that public sector organizations should respect privacy and data protection regulations, in order to ensure the benefit for citizens.

The selection of these cases is based on the fact that they can be classified under North American and European AI models, respectively. However, both countries share common history and language, as well as administrative cultures. Therefore, this is an opportunity to compare and understand different perspectives and visions about the governance of AI. Hence, the research design is oriented to describe one case of the North American model of AI governance and one case of the European AI governance model. Our study compares these two cases and the opinions of top IT public officials in ministries of both countries, Mexico and Spain.

Public managers’ perceptions and AI policies in the public sector

Given the nascent nature of AI and algorithms in the public sector, many governments are accelerating expectations about their implications. At the same time, they are working on understanding this new wave of technological evolution, their applications, and impact. Furthermore, scholars do not agree about the implications of AI in the public sector. This ambiguity is even more important among public officials and senior managers in public organizations.

This study emerges from the fact that the role of CIOs is critical to understanding the design and utilization of technologies in the public sector. Their role is particularly important at this early stage of AI adoption, a moment of transition from traditional digital government applications to algorithmic-based technologies. The role of CIOs had a disruptive growth in the 1980s as a consequence of a worldwide digital transformation (Haffke, et al., 2016; Weill and Woerner, 2013). The role of CIO is defined as the highest level of IT executive or manager in a firm or business unit (Banker, et al., 2011). In the case of the public sector, CIOs also have gained a predominant role in translating political strategies into administrative and service plans and actions. In other words, the growing role of the so-called infocracy (Bovens and Zouridis, 2002; Zuurmond, 1994) is led by CIOs in the public sector, who are essential in the initial moments of diffusion of innovations and technologies, as they frame, advocate, and monitor the implementation of projects within their ministerial departments.

Therefore, this study is intended to provide empirical evidence derived from a comparative analysis of perceptions, expectations, and challenges/opportunities about AI of top public officials, in two different countries. In both cases, this group of CIOs are dealing with IT and AI in federal/national contexts, within different policy areas (ministries). Particularly, the role of top CIOs is key as they play a leading role within each ministry, and they are responsible of the IT personnel and processes within their governments. Hence, this article illuminates the debate about AI in the public sector comparing perceptions and expectations of those who are managing technologies and AI in ministries of two different national governments, and the challenges/opportunities that they identify in the process of adoption and implementation of AI in public settings.

That being said, there are some differences regarding CIOs as a professional group of public managers that need to be underlined. In Spain, the CIO leading IT in each ministry is part of a group of technological public sector managers (corp), named “Information technology and systems group” (Cuerpo superior de sistemas y tecnologías de la información), established in 1990. This infocratic group of selected public sector managers, with a technological profile, maintain their positions independent of changes in national government, while keeping important levels of autonomy from political appointees, all to develop technological strategies and policies. Hence, Spain exhibits a stable group of IT professionals who are recruited via open examination (oposición) and work permanently in different government ministries and departments leading main technological decisions. On the other hand, in Mexico CIOs are openly appointed by the ministries or maintain personnel that was working in a previous administration. This is completely at the discretion of the minister himself or herself. Thus, in this case, senior IT officials enjoy less stability and more transferability between private and public sectors.

 

++++++++++

Analytical strategy and research methods

This study analyzes artificial Intelligence (AI) momentum in national/federal ministries of two different countries. At this stage, national governments are mostly responsible of regulation, promotion, and management of AI and algorithm-mediated governance initiatives in a majority of countries. Moreover, the objective of this research is understanding the perceptions of several CIOs regarding AI at this layer of government. Specifically, we are examining the following research questions:

RQ1: What is the perception of those who manage AI about this concept?
RQ2: What are the expectations of those who manage AI regarding its development in public administration?
RQ3: What are the main challenges and opportunities for public administrations managing AI?

This exploratory research is based on primary data about the perceptions of top IT public officials within the Mexican and Spanish ministries. The aim of this study is to examine the institutional perceptions of national governments, looking at opinions of CIOs responsible for managing AI. Up to now, research about AI in the public sector is mostly based on normative approaches or case studies, with very limited quantitative and comparative data. Literature on AI in the public sector does not give voice to the perceptions of IT managers responsible for policy at a national level of government. Consequently, this article offers key data based on a survey pinpointing the perceptions of public officials designing AI nationally.

Analytical strategy

This research seeks to identify institutional perspectives of federal/national governments regarding different dimensions of AI in public administration. Our study includes the following dimensions, based on opinions of senior public officials managing technologies:

  1. Perceptions regarding to the concept of AI in the public sector. Currently, the notion of AI is not completely shared by scholars and practitioners working in the public sector. Our assumption is that CIOs will be closer to technocentric visions and images of AI, including big data analytics, robotics, and automated systems, than a citizen-centric or social side of disruptive technologies in the public sector.

  2. Expectations about the development of AI in the public sector. Here, our surmise is that CIOs will not have transformative prospects about the implementation of AI in the public sector; they will share techno-optimistic expectations about the future. As it has been identified in previous studies, one may expect that they will share optimistic ideas regarding the potential of AI in the public sector, oriented to foster traditional gains in public management operations (i.e., efficiency, digitalization, or transparency), instead of more disruptive (and political) changes in government, especially in relation of public sector institutions to citizenry.

  3. Challenges and opportunities of AI in the public sector. This dimension studies how public managers demonstrate the realities of AI implementation in the public sector. Thus, we analyze factors that are discouraging and promoting the diffusion of AI in the public sector. Here, our assumption is that AI techniques and applications are in its infancy and CIOs are facing inhibitors in the process of implementation, as well as opportunities. Nonetheless, they do not have a clear idea of the real implications in the public sector.

Research methods

We administered an online survey to federal/national government CIOs leading ministerial policies in two different countries, Mexico and Spain. The names and contact data of the CIOs were compiled from Web sites of their corresponding ministries and departments. The contact information, including e-mail addresses of the national government CIOs, was verified for accuracy by direct communication with their offices over the phone. All the ministries were e-mailed a cover letter requesting their participation in the survey, including a Web link to the online instrument. The anonymous treatment of the answers was one of the concerns of the CIOs and their offices. Anonymity was guaranteed in the cover letter, presenting the purpose and motivation of this study, as well as descriptions of the researchers and affiliations involved in this international project.

The research technique to collect data comprised a survey (see Appendix to review analytical strategy and survey questions). The questionnaire consisted in 19 questions aimed at public officials managing ICTs in the ministries of Mexico and Spain. The period to receive answers to the survey was opened from 15 June to 15 August 2019. The survey obtained a high level of commitment of the high-level public professionals; the response rate was almost 74 percent of the ministries surveyed (11 of 18, 61 percent in Mexico, and 15 of 17, 88 percent in Spain). The validation of this survey was supported by reviews of public managers in both countries who examined the content to avoid inconsistencies.

The survey was designed and constructed based on a diverse literature review focused on empirical and theoretical studies, such as Pombo, et al. (2018), Russell and Norvig (2016), Tinholt, et al. (2017), Tito (2017), and Zhang and Dafoe (2019). These investigations identified various typologies and concepts that were very useful for creating the final questions.

In order to examine the dimensions of our study, we analyzed survey results using descriptive statistics. These rather simple but effective methods were appropriate to reach our exploratory research goal. Regarding the small survey sample, each federal/national government is formed by 18 ministries, each of them including one corresponding CIO. Therefore, the analysis conducted in this exploratory study is based on the answers of almost all of the Mexican and Spanish ministries.

 

++++++++++

Results

This section provides results on the three main dimensions that we surveyed of public officials: perceptions, expectations, and opportunities/challenges of AI in government. Our study provides primary data from CIOs leading technological policies in national/federal governments.

Table 1 illustrates demographic information on responding CIOs in both countries. The medium age of respondents ranges from 49.3 in the case of Spain to 43.7 in Mexico. The vast majority of CIOs were male, representing more than 80 percent of the respondents. In both countries, the typical academic degree of CIOs was bachelor’s, whereas more than 30 percent in Spain and 40 percent in Mexico had a master’s or doctorate. In almost all cases CIOs studied informatics and telecommunications engineering, as well as mathematics, physics, and statistics. These demographics presented a group of CIOs relatively young, predominantly male, and with advanced and diverse technological educational backgrounds.

 

Table 1: Information on national/federal CIOs in Mexico and Spain.
 SpainMexico
Ministries15 (17)11 (18)
Average age49.343.7
Gender  
Female13.3%18.2%
Male86.7%81.8%
Academic degrees  
Bachelor’s66.7%54.5%
Master’s20.0%27.3%
Doctorate13.3%18.2%

 

Perceptions about the concept of AI in the public sector

In the first instance, our study examines the perceptions of CIOs about AI in the public sector, as emergent technologies are not unanimously perceived, adopted, and implemented. In particular, the perceptions of senior managers will help to recognize the scope and potential of AI in the public sector as this group of public managers “frame” IT policies. Hence, this part of the study sheds light on the concepts linked, and techniques most commonly associated, to AI by national/federal CIOs.

The analysis of the overarching question about AI shows some interesting results. The survey question here was, “to what extent do you agree with the following ideas regarding to artificial intelligence”. The responses for the question in the survey were on a seven-point Likert scale, ranging from “totally agree” (7) to “totally disagree” (1). The results are summarized in Table 2, including average scores and standard deviations in both countries. CIOs in Spain seem more open to the adoption of AI in the public sector (6.7), than their colleagues in Mexico (4.9). Also, this is the case of the future demand of other “intelligences” in the public sector (5.7 in Spain, 4.6 in Mexico). Regarding the future of jobs, IT managers, more (in Spain, 5.7) or less (in Mexico, 4.7), assume naturally that robots and humans will share jobs, and that new professions in the public sector linked to AI will have to be addressed thoroughly (again, in Spain: 6.3, in Mexico: 4.6). In both cases, there exist a common approach to the complexity of envisioning future occupations of public employees, as the replacement process will have profound effects (5.3./5.2). Likewise, they almost agree that AI in the public sector is not very different than AI in the private sector (Spain: 5.3; Mexico: 4.6).

 

Table 2: General ideas related to artificial intelligence in the public sector.
 Spain
(n=15)
Mexico
(n=11)
AverageStandard deviationAverageStandard deviation
I am totally open to the adoption of AI.6.70.54.92.5
AI in the public sector is not very different than AI in the private sector.5.31.84.62.2
I assume naturally that robots and humans will share jobs. 5.71.34.72.1
Envisioning future occupations of public employees is complex, as the replacement process will have profound impacts.5.31.75.21.8
Other “intelligences” will be required for public employees.5.71.64.61.9
The new professions in the public sector linked to AI will have to be tackled thoroughly.6.31.64.61.6
Average5.81.44.82.0

 

Finally, a dimension included a category oriented to identify human capacities and behaviors that CIOs link with transformations based on the application of AI in the public sector. The survey question about capacities/behaviors was, “regarding to this classification of human capacities and behaviors, to what extent do you agree with their replacement by artificial intelligence in the public sector?” The responses for the question in the survey were on a seven-point Likert scale, ranging from “totally agree” (7) to “totally disagree” (1). In Table 3, CIOs in both countries report common experiences about the transformation of capacities and behaviors based on AI in the public sector. In both countries, human capacities and behaviors sharing the highest transformative scores are monitor (6.2/6.3), anticipate (5.8/6.3), remember (6.0/5.7), and analyze (4.9/5.9). Also, in both countries CIOs identify two categories in the other side of the continuum: feel (3.1/2.7) and moralize (3.2/2.3), the human capacities/behaviors more reluctant to be transformed by AI.

 

Table 3: Human capacities and behaviors based on artificial intelligence.
 Spain
(n=15)
Mexico
(n=11)
AverageStandard deviationAverageStandard deviation
Monitor6.30.76.21.5
Monitor6.30.76.21.5
Analyze5.91.04.92.3
Act4.12.14.02.0
Interact4.91.54.02.4
Remember5.71.26.01.6
Anticipate6.30.95.81.9
Feel 2.71.93.11.7
Moralize2.31.83.21.8
Create3.71.94.32.3
Decide4.31.84.02.0
Average4.61.54.52.0

 

Expectations about AI in the public sector

Social and political expectations about AI are high in terms of achievement and benefits, but also potential pitfalls and problems. This dimension comprised expectations of CIOs about transformational capacities of AI in the public sector. This part of the study assessed the expected impact of AI in different administrative functions in the public sector. In order to understand this dimension, the survey question was, “from your point of view, what of the following functions will be more affected by artificial intelligence in the short term?” The responses in the survey were to a multi-answer question (with a minimum of three options to answer from the complete list, see Figure 1). In the opinion of CIOs, the most affected functions by the implementation of AI will be clerical and assistant activities (63.6 in Mexico and 46.7 in Spain), and processing of transactions/operations (63.6 in Mexico and 93.3 in Spain). Conversely, executive management (9.1 in Mexico and 13.3 in Spain) and regulation (13.3 in Mexico and 36.4 in Spain), were the functions that CIOs expected to be less altered by AI-based techniques. Confirming previous studies, they saw repetitive and low added-value functions most predisposed to be automated with AI, whereas those involving creative processes and tasks would remain unchanged.

 

Expected functions more affected by the implementation of artificial intelligence
 
Figure 1: Expected functions more affected by the implementation of artificial intelligence.

 

The final category of this dimension evaluates the expected impact of AI in different public policy domains of the executive branch. In order to understand this dimension, the survey question was, “from your point of view, in which public policy domains will be adopted artificial intelligence at an earliest stage?” The responses in the survey were on a multi-answer question (with a minimum of three options to answer from the complete list; see Figure 2). Here, some differences between appeared between each country. In the case of Mexican CIOs, only the areas of security (63.3 percent) and mobility (54.5 percent) reach 50 percent, expecting early adoption. On the contrary, Spanish CIOs described more policy areas that will be adopting AI, including mobility (73.3 percent), and security (53.3 percent), but also defense (60.0 percent), tourism (60.0 percent), and public health (53.3 percent), all these including over 50 percent of respondents.

 

Expected policy areas adopting artificial intelligence at an early stage
 
Figure 2: Expected policy areas adopting artificial intelligence at an early stage.

 

Challenges and opportunities of AI in the public sector

This final dimension of our study described challenges and opportunities based on realities in the application of AI in the public sector. Here, attention was given to IT leaders in ministries promoting or planning, at least in some extent, applications, cases, and techniques, based on algorithms, big data analytics, and AI technologies, systems, or applications, facing different inhibitors and enablers.

First, analyzing inhibitors defined how CIOs understand real challenges to the implementation of AI in their organizations. In order to understand this dimension, the survey question was, “what are the main inhibitors for the implementation of artificial intelligence in the public sector?” The responses in the survey were to a multi-answer question (with a minimum of three options to answer from the complete list). Figure 3 illustrates that, in both cases, there are three key inhibitors, including budget (90.9 percent), digital divide (63.6 percent), and technological infrastructure (54.5 percent) all in Mexico, and legislation (73.3 percent), budget (60.0 percent), and administrative culture (53.3 percent) in Spain. Budgetary concerns were a key issue, and actual legislation about personal data protection and administrative procedures was also an actual concern in the Spanish case. In Mexico, it was also remarkable that aspects concerning security and privacy of data were not identified as impediments for the implementation of AI in the public sector (0 percent of respondents chose these options), corresponding with a North American approach to AI. This category reports real differences concerning implementation experiences in both cases. At the same time, inhibitors referred to distinctive problems in the public sector, advancing, in some extent, singular perspectives in different AI governance systems.

 

Inhibitors of artificial intelligence in public organizations
 
Figure 3: Inhibitors of artificial intelligence in public organizations.

 

Studying enablers provided insights into how CIOs understand actual opportunities in the implementation of AI in their organizations. With respect to facilitators, the survey question was, “what are the main enablers for the implementation of artificial intelligence in the public sector?” The responses in the survey were on a multi-answer question (with a minimum of three options to answer from the complete list; see Figure 4). In this case, the opinions of CIOs were very diverse, not concentrated in specific options. For Spain, only the items knowledge (73.3 percent) and technological infrastructure (53.3 percent) exceed 50 percent of responses. In Mexico, CEOs appeared less knowledgeable over AI implementation in their organizations, as only technological transfer (46.7 percent) collected a significant percentage of answers. This question on enablers provided evidence about different experiences and results in the nascent process of implementation of AI in Spain and Mexico.

 

Enablers of artificial intelligence in public organizations
 
Figure 4: Enablers of artificial intelligence in public organizations.

 

 

++++++++++

Discussion

This article reported on answers to three research questions based on the nascent, and evolving, state of the art of AI in government: (1) What is the perception of those who manage AI about this concept? (Perceptions regarding the concept of AI in the public sector); (2) What are the expectations of those who manage AI regarding its development in public administration? (Expectations about the development of AI in the public sector); (3) What are the main challenges and opportunities for public administrations managing AI? This exploratory study presented the results of this survey administered to federal/national government CIOs in Mexico and Spain. Our descriptive statistical analysis provided an overall analysis. At the same time, our primary data is a resource about AI in the public sector, encouraging future research as well as informing national AI policies and strategies.

Our first question about perceptions regarding AI in the public sector generated answers from CIOs that displayed high level of similarities between both countries. CIOs in Mexico and Spain shared common beliefs about AI, including the future of jobs in the public sector, as well as the generation of new professions in the public sector linked to AI. CIOs almost all agreed that AI in the public sector was not very different than AI in the private sector and, in general, they were very open about adopting AI to public sector settings. Similarly, they provided more transformative options to some human capacities and behaviors (monitor, anticipate, remember, and analyze), than to others (feel or moralize). Likewise, these results confirmed the existence of common views about AI in governments and public administration in both countries. Besides, they supported our surmise about CIOs, who assign a technocentric nature to AI in the public sector, leaving aside social/human-centric approaches.

We then analyzed CIOs’ expectations about the development of AI in the public sector in the future. Noticeable differences have been reported about these transformational capacities of AI in the public sector. These differences between Mexican and Spanish CIOs were very visible regarding public policy domains that, in their opinion, will be more affected by AI at an early stage. For Mexico, security and mobility were pinpointed, while for Spain, mobility, and security were targeted, as well as defense, tourism, and public health. Consequently, these results corroborate different expectations over the potential of AI in the public sector. At the same time, these results confirmed our initial assumption that CIOs expect mainly positive outcomes, more oriented to foster public management operations (more accentuated in Mexico), than other probably more crucial areas of activity in government and public administration (more emphasized in Spain).

We then examined challenges and opportunities based on realities of implementation of AI in public administration. In this category both groups of CIOs displayed dissimilarities in their understanding, on the one hand, of inhibitors of AI implementation. In the Mexican case, they reported digital divide and technological infrastructure (with a remarkable lack of attention to the security and privacy of data). In Spain, CIOs pointed out legislation and administrative culture, sharing a common concern with budget. These differences appeared to address enablers of AI implementation, with CIOs in Mexico less experienced than those in Spain. As it was initially expressed, neither one group of IT managers nor the other had developed yet a well-defined prospect of inhibitors and enablers of AI implementation in the public sector. Among other options, this confirmed that we are immersed in a first stage of applications, cases, and systems based on AI in the public sector. A lack of experience was noticeable even among senior IT managers.

The implications of this study are manifold. At first sight, our study highlighted an apparent contradiction with our initial purpose: CIOs seem to share a common understanding of AI in the public sector and, most surprisingly, the idea that AI in government is not very different than in the private sector. This is an intriguing conclusion of our study that needs further investigation. On the one hand, public administration scholars are looking for the singularity of public agencies implementing IA (Sun and Medaglia, 2019; Valle-Cruz, et al., 2019). On the other, those who are involved in that process seem to be reluctant to accept differences with the private sector. Then, one may point out some ideas to support our initial goal of achieving an exclusive public sector approach to AI (Margetts and Dorobantu, 2019). They include ethical issues derived from the application of biased algorithms, social responsibility of public organizations to deliver big data analysis based on safe and equal treatment of citizens (above all, in sensitive areas like social care, health, and education) (Ananny, 2016), or the need to value the governance of historical and contextual knowledge of individuals and organizations.

Looking at the theoretical realm, complementary issues emerged from our analysis of AI in the public sector. Some have claimed that the fourth technological revolution is in process while others suggested that AI is not a silver bullet (Vydra and Klievink, 2019). Therefore, in order to assess this process, we needed to use theories and approaches rooted in social sciences, asking among other questions, why some governments/countries/regions are more open to implement AI policies and strategies, while others are more skeptical; how AI is oriented to foster some policy domains and managerial functions, and not others; or to what extent governments and public administrations use AI to strengthen, rather than destroy, democratic values and citizen-centric governance models in times of uncertainty, populism and crisis of representative democracies. There are several perspectives to understand the results, including policy diffusion (Rogers, 1995) or technology enactment (Fountain, 2001; Gil-Garcia, 2012; Orlikowski, 2000) in government, as they will play a role in the near future. At the same time, the transformative and performative nature of algorithms and AI techniques may challenge our understanding of data analysis and methods.

Precisely, our results shed light on the coincidences and discrepancies that CIOs of federal/national governments in Spain and Mexico have in relation to AI in the public sector. Our survey respondents represented key decision-makers in their countries. We reported noticeable discrepancies looking at key issues, including expectations about the future, and actual realities of implementation. According to the literature, different AI governance models are in play at this moment (Lee, 2018; Wright, 2018). Our data suggest that our selected cases are exemplary, at least in some extent, of North American (Mexico) and European (Spain) models. In Mexico, AI in the public sector appeared to be more business-driven, while data privacy and protection were less important for politicians and public institutions. In Spain, AI in the public sector seemed to be more citizen-centric, and data protection and privacy represented real concerns for politicians and public institutions. Working with the hypothesis of different models of AI governance in the public sector deserves attention of research in coming years. In our opinion, not only the future of public administration is at stake, but also the very nature of our political and democratic systems.

Also, looking at the role of CIOs was another key issue as top infocracy members will shape emergent AI policies and strategies. Here, we foresee that future public administrators will have to establish multidisciplinary work teams to respond to public challenges and problems. The technological revolution that AI brings with it is showing a trend in the public sector, oriented towards a convergence between political, administrative, social sciences and data, information systems, and computer sciences. Although the CIOs of both national governments were technological positivists, the issue of ethical values with the use of AI was supported by the results of this study. Senior IT managers had reservations over the implementation of AI in government, because they considered possible biases in social values and ethical principles.

Based on our findings, we may predict all levels of government need to strengthen their analytical capabilities associated with the implementation of policies in different areas. This implies developing analytical strategies such as policy modeling, big data analysis, or cloud computing, but also operational understanding of contextual settings in order to cope complex societal problems.

 

++++++++++

Conclusion

This study was an attempt to understand first steps of artificial intelligence (AI) in governments. Based on the opinions of CIOs, one of the main inferences was that each of these countries is divergently promoting AI, at least in some extent, not only in government itself, but also from government to the society (fostering business-centric or citizen-centric perspectives, giving more or less prominence to data protection, privacy, and security, or prioritizing some functions and policy sectors and not others), and also designing different governance models.

This effort examined differences to public and private sectors related to AI implementation, emerging threats for democratic values and representative systems based on algorithm-based governance, growing roles of infocracy in the public sector, or the challenging patterns of future research based on big data analytics and algorithms in public administration.

On the practical side, this research provided information to public managers interested in the implications and critical issues regarding AI in governments and public administrations. In particular, this article fosters an interest in policy areas with more potential to implement AI applications and techniques; managerial functions that are more open to the adoption of AI applications and techniques; and challenges/risks that public managers may face to define AI policies and projects in the public sector.

There are limitations to this study. On the one hand, our study is based on national/federal CIOs of Mexico and Spain, so our conclusions cannot be directly inferred to other different layers of government. We conducted research about a selected group of CIOs leading each ministry, providing a small sample. Additionally, this research is exploratory, based on descriptive statistical analysis. We hope that this study will encourage future research in other countries and administered within a more diverse universe of IT public managers. Additionally, more refined statistical techniques would facilitate explanatory conclusions in future studies.

In sum, the future of AI in governments and public administrations is not foreordained. Different voices are claiming for fair, anti-biased, and equity-based AI policies and strategies. In the case of the public sector, the responsibility of political appointees and senior managers for ethical issues is even greater than in the private sector, regarding the implications of applications, cases, and projects for the public. In our opinion, research about AI also needs to tackle a growing complexity and variety of datasets, analytical methods, and devices to gather information from the society, companies, and public sector organizations. At the end of the day, this will be in parallel with our efforts analyzing collective action, decision-making processes, and the management of societal problems in government institutions. Correspondingly, expected deep implications of AI in our democratic values and political systems invite us to be extremely cautious with inferences derived from studies of this emerging area. End of article

 

About the authors

J. Ignacio Criado (Ph.D.) is an Associate Professor/Senior Lecturer in Political Science and Public Administration, Department of Political Science and International Relations, at Universidad Autónoma de Madrid. His articles have been published in leading journals, including Government Information Quarterly, Social Science Computer Review, First Monday, Information Polity, Local Government Studies, International Journal of Public Administration, and International Journal of Public Sector Management (among others). His research interests include algorithmic governance and artificial intelligence in the public sector; open government and policies for transparency, participation and public innovation; and, social media and big data in government.
E-mail: ignacio [dot] criado [at] uam [dot] es

Rodrigo Sandoval-Almazan (Ph.D.) is Associate Professor at the Political Sciences and Social Sciences Department in the Autonomous University of the State of Mexico, in Toluca City. Dr. Sandoval-Almazan is the author or co-author of articles in Government Information Quarterly, Information Polity, First Monday, Government; Journal of Information Technology for Development; Journal of Organizational Computing and Electronic Commerce; and, International Journal of E-Politics IJEP. His research interests includes: artificial intelligence, social media in government. public innovation, digital government, and open government.
E-mail: rsandovala [at] uaemex [dot] mx

David Valle-Cruz (Ph.D.) is Professor at the Computing Engineering Department in the Autonomous University of the State of Mexico, in Toluca City. He is the author or co-author of articles in Government Information Quarterly, Information Polity, First Monday, International Journal of Public Sector Management, International Journal of Public Administration in the Digital Age, and Digital Government: Research and Practice. His research is related to applied artificial intelligence and emerging technologies in government.
E-mail: davacr [at] uaemex [dot] mx

Edgar Alejandro Ruvalcaba-Gómez (Ph.D.) is a Research Professor at the Universidad de Guadalajara, with affiliation to the Department of Public Policy, and Coordinator in the Research Institute in Public Policies and Government. He holds a Ph.D. in Law, Government and Public Policy from Autonomous University of Madrid, Spain. Edgar conducts research related to open government, transparency, citizen participation, open data, corruption, digital government and artificial intelligence, and public innovation.
E-mail: edgar [dot] ruvalcaba [at] cucea [dot] udg [dot] mx

 

Acknowledgements

This study was partially supported by the Research Grant H2019-HUM 5699 (On Trust), Madrid Regional Research Agency and European Social Fund.

 

References

P.K. Agarwal, 2018. “Public administration challenges in the world of AI and bots,” Public Administration Review, volume 78, number 6, pp. 917–921.
doi: https://doi.org/10.1111/puar.12979, accessed 26 December 2020.

T. Ahmad, R. Aljafari, and V. Venkatesh, 2019. “The government of Jamaica’s electronic procurement system: Experiences and lessons learned,” Internet Research, volume 29, number 6, pp. 1,571–1,588.
doi: https://doi.org/10.1108/INTR-02-2019-0044, accessed 26 December 2020.

M. Ananny, 2016. “Toward an ethics of algorithms: Convening, observation, probability, and timeliness,” Science, Technology, & Human Values, volume 41, number 1, pp. 93–117.
doi: https://doi.org/10.1177/0162243915606523, accessed 26 December 2020.

D. Araya, 2019. “Who will lead in the age Of artificial intelligence?” Forbes (1 January), at https://www.forbes.com/sites/danielaraya/2019/01/01/who-will-lead-in-the-age-of-artificial-intelligence/, accessed 26 December 2020.

R.D. Banker, N. Hu, P.A. Pavlou, and J. Luftman, 2011. “CIO reporting structure, strategic positioning, and firm performance,” MIS Quarterly, volume 35, number 2, pp. 487–504.
doi: https://doi.org/10.2307/23044053, accessed 26 December 2020.

T.J. Barth and E. Arnold, 1999. “Artificial intelligence and administrative discretion: Implications for public administration,” American Review of Public Administration, volume 29, number 4, pp. 332–351.
doi: https://doi.org/10.1177/02750749922064463, accessed 26 December 2020.

M. Bovens and S. Zouridis. 2002. “From street-level to system-level bureaucracies: How information and communication technology is transforming administrative discretion and constitutional control,” Public Administration Review, volume 62, number 2, pp. 174–184.
doi: https://doi.org/10.1111/0033-3352.00168, accessed 26 December 2020.

E. Brynjolfsson and A. McAfee, 2017. “The business of artificial intelligence,” Harvard Business Review (18 July), at https://hbr.org/2017/07/the-business-of-artificial-intelligence, accessed 26 December 2020.

E. Brynjolfsson and T. Mitchell, 2017. “What can machine learning do? Workforce implications,” Science, volume 358, number 6370 (22 December), pp. 1,530–1,534.
doi: https://doi.org/10.1126/science.aap8062, accessed 26 December 2020.

T. Bucher, 2017. “The algorithmic imaginary: Exploring the ordinary affects of Facebook algorithms,” Information, Communication & Society, volume 20, number 1, pp. 30–44.
doi: https://doi.org/10.1080/1369118X.2016.1154086, accessed 26 December 2020.

J. Burrell, 2016. “How the machine ‘thinks’: Understanding opacity in machine learning algorithms,” Big Data & Society.
doi: https://doi.org/10.1177/2053951715622512, accessed 26 December 2020.

S. Clarkson, 2008. Does North America exist? Governing the continent after NAFTA and 9/11. Toronto: University of Toronto Press.

J.I. Criado and J.R. Gil-Garcia, 2019. “Creating public value through smart technologies and strategies: From digital services to artificial intelligence and beyond,” International Journal of Public Sector Management, volume 32, number 5, at https://www.emerald.com/insight/content/doi/10.1108/IJPSM-07-2019-0178/full/html, accessed 26 December 2020.

J.I. Criado, J. Valero, and J. Villodre, 2020. “Artificial intelligence and decision-making discretion: An analysis of SALER algorithmic early warning system,” Information Polity, volume 25, number 4, pp. 449–470.
doi: https://doi.org/10.3233/IP-200260, accessed 26 December 2020.

J.I. Criado, R. Sandoval-Almazan, and J.R. Gil-Garcia, 2013. “Government innovation through social media,” Government Information Quarterly, volume 30, number 4, pp. 319–326.
doi: https://doi.org/10.1016/j.giq.2013.10.003, accessed 26 December 2020.

K. Cukier and V. Mayer-Schoenberger, 2013. “The rise of big data: How it’s changing the way we think about the world,” Foreign Affairs, volume 92, number 3, pp. 28–40, and at https://www.foreignaffairs.com/articles/2013-04-03/rise-big-data, accessed 26 December 2020.

K.C. Desouza, 2018. “Delivering artificial intelligence in government: Challenges and opportunities,” IBM Center for the Business of Government, at http://www.businessofgovernment.org/report/delivering-artificial-intelligence-government-challenges-and-opportunities, accessed 26 December 2020.

T. Dutton, 2018. “An overview of national AI strategies” (28 June), at https://medium.com/politics-ai/an-overview-of-national-ai-strategies-2a70ec6edfd, accessed 26 December 2020.

European Commission, 2019. “eGovernment benchmark 2019. Trust in government is increasingly important for people” (18 October), at https://ec.europa.eu/digital-single-market/en/news/egovernment-benchmark-2019-trust-government-increasingly-important-people, accessed 26 December 2020.

European Commission, 2018. “Artificial intelligence: A European perspective,” at https://ec.europa.eu/jrc/en/publication/artificial-intelligence-european-perspective, accessed 26 December 2020.

European Commission, Joint Research Center, 2020. “AI watch — Artificial intelligence in public services,” at https://ec.europa.eu/jrc/en/publication/eur-scientific-and-technical-research-reports/ai-watch-artificial-intelligence-public-services, accessed 26 December 2020.

J.E. Fountain, 2001. Building the virtual state: Information technology and institutional change. Washington, D.C.: Brooking Institution Press.

S. Ganapati and C.G. Reddick, 2012. “Open e-government in U.S. state governments: Survey evidence from chief information officers,” Government Information Quarterly, volume 29, number 2, pp. 115–122.
doi: https://doi.org/10.1016/j.giq.2011.09.006, accessed 26 December 2020.

D. Gershgorn, 2019. “California police are sharing facial recognition databases to ID suspects” (1 August), at https://onezero.medium.com/california-police-are-sharing-facial-recognition-databases-to-id-suspects-3317726d31ad, accessed 26 December 2020.

S. Giest, 2017. “Big data for policy-making: Fad or fasttrack?” Policy Sciences, volume 50, pp. 367–382.
doi: https://doi.org/10.1007/s11077-017-9293-1, accessed 26 December 2020.

J.R. Gil-Garcia, 2012. Enacting electronic government success: An integrative study of government-wide Websites, organizational capabilities, and institutions. New York: Springer.
doi: https://doi.org/10.1007/978-1-4614-2015-6, accessed 26 December 2020.

J.R. Gil-Garcia, S.S. Dawes, and T.A. Pardo, 2018. “Digital government and public management research: Finding the crossroads,” Public Management Review, volume 20, number 5, pp. 633–646.
doi: https://doi.org/10.1080/14719037.2017.1327181, accessed 26 December 2020.

I. Haffke, B.J. Kalgovas, and A. Benlian, 2016. “The role of the CIO and the CDO in an organization’s digital transformation,” Thirty-Seventh International Conference on Information Systems, at https://aisel.aisnet.org/icis2016/ISStrategy/Presentations/3/, accessed 26 December 2020.

L.D. Introna, 2015. “Algorithms, governance, and governmentality: On governing academic writing,” Science, Technology, & Human Values, volume 41, number 1, pp. 17–49.
doi: https://doi.org/10.1177/0162243915587360, accessed 26 December 2020.

M. Janssen and G. Kuk, 2016. “The challenges and limits of big data algorithms in technocratic governance,” Government Information Quarterly, volume 33, number 3, pp. 371–377.
doi: https://doi.org/10.1016/j.giq.2016.08.011, accessed 26 December 2020.

M. Janssen, S. Luthra, S. Mangla, N.P. Rana, and Y.K. Dwivedi, 2019. “Challenges for adopting and implementing IoT in smart cities: An integrated MICMAC-ISM approach,” Internet Research, volume 29, number 6, pp. 1,589–1,616.
doi: https://doi.org/10.1108/INTR-06-2018-0252, accessed 26 December 2020.

D. Lazer, 2016. “The rise of the social algorithm,” Science, volume 348, number 6239 (7 May), pp. 1,090–1,091.
doi: https://doi.org/10.1126/science.aab1422, accessed 26 December 2020.

K.-F.Lee, 2018. AI Super-powers. China: Silicon Valley, and the new world order. New York: Houghton Mifflin Harcourt.

L.F. Luna-Reyes, S. Mellouli, J.R. Gil-Garcia, T.A. Pardo, J. Zhang, and C. Navarrete, 2010a. “Comparing digital government agendas in Canada, Mexico, and the United States,” dg.o ’10: Proceedings of the 11th Annual International Digital Government Research Conference on Public Administration Online: Challenges and Opportunities, pp. 241–242.

L.F. Luna-Reyes, T.A. Pardo, J.R. Gil-Garcia, C. Navarrete, J. Zhang, and S. Mellouli, 2010b. “Digital government in North America: A comparative analysis of policy and program priorities in Canada, Mexico, and the United States,” In: C.G. Reddick (editor). Comparative e-government. New York: Springer, pp. 139–160.
doi: https://doi.org/10.1007/978-1-4419-6536-3_7, accessed 26 December 2020.

H. Margetts, 2017. “The data science of politics,” Political Studies Review, volume 15, number 2, pp. 201–209.
doi: https://doi.org/10.1177/1478929917693643, accessed 26 December 2020.

H. Margetts and C. Dorobantu, 2019. “Rethink government with AI,” Nature, volume 568, number 7751 (11 April), pp. 163–165.
doi: https://doi.org/10.1038/d41586-019-01099-5, accessed 26 December 2020.

A. Meijer, 2018. “Datapolis: A public governance perspective on ‘smart cities’,” Perspectives on Public Management and Governance, volume 1, number 3, pp. 195–206.
doi: https://doi.org/10.1093/ppmgov/gvx017, accessed 26 December 2020.

A. Meijer, M. Lips, and K. Chen, 2019. “Open governance: A new paradigm for understanding urban governance in an information age,” Frontiers in Sustainable Cities (28 August).
doi: https://doi.org/10.3389/frsc.2019.00003, accessed 26 December 2020.

A. Meijer, J.R. Gil-Garcia, and M.P. Rodríguez-Bolvar, 2016. “Smart city research: Contextual conditions, governance models, and public value assessment,” Social Science Computer Review, volume 34, number 6, pp. 647–656.
doi: https://doi.org/10.1177/0894439315618890, accessed 26 December 2020.

I. Mergel, R.K. Rethemeyer, and K. Isett, 2016. “Big data in public affairs,” Public Administration Review, volume 76, number 6, pp. 928–937.
doi: https://doi.org/10.1111/puar.12625, accessed 26 December 2020.

W.J. Orlikowski, 2000. “Using technology and constituting structures: A practice lens for studying technology in organizations,” Organization Science, volume 11, number 4, pp. 367–472.
doi: https://doi.org/10.1287/orsc.11.4.404.14600, accessed 26 December 2020.

I. Pencheva, M. Esteve, and S.J. Mikhaylov, 2020. “Big data and AI — A transformational shift for government: So, what next for research?” Public Policy and Administration, volume 35, number 1, pp. 24–44.
doi: https://doi.org/10.1177/0952076718780537, accessed 26 December 2020.

C. Pombo, R. Gupta, and M. Stankovic, 2018. “Social services for digital citizens: Opportunities for Latin America and the Caribbean,” Inter-American Development Bank.
doi: http://dx.doi.org/10.18235/0001105, accessed 26 December 2020.

E.M. Rogers, 1995. Diffusion of innovations. Fourth edition. New York: Free Press.

S.J. Russell and P. Norvig, 2016. Artificial intelligence: A modern approach. Third edition. Upper Saddle River, N.J.: Pearson.

H.A. Sapp, 2005. “North American anti-circumvention: Implementation of the WIPO Internet treaties in the United States, Mexico and Canada,” Computer Law Review and Technology Journal, volume 10.

N. Seaver, 2013. “Knowing algorithms,” at http://nickseaver.net/writing, accessed 26 December 2020.

M. Sideri, A. Kitsiou, A. Filippopoulou, C. Kalloniatis, and S. Gritzalis, 2019. “E-governance in educational settings: Greek educational organizations leaderships perspectives towards social media usage for participatory decision-making,” Internet Research, volume 29, number 4, pp. 818–845.
doi: https://doi.org/10.1108/IntR-05-2017-0178, accessed 26 December 2020.

Spain. Ministry of Science and Universities, 2019. “Spanish RDI strategy in artificial intelligence,” at https://www.ciencia.gob.es/stfls/MICINN/Ciencia/Ficheros/Estrategia_Inteligencia_Artificial_EN.PDF, accessed 26 December 2020.

T.Q. Sun and R. Medaglia, 2019. “Mapping the challenges of artificial intelligence in the public sector: Evidence from public healthcare,” Government Information Quarterly, volume 36, number 2, pp. 368–383.
doi: https://doi.org/10.1016/j.giq.2018.09.008, accessed 26 December 2020.

D. Tinholt, S. Enzerink, W. Sträter, P. Hautvast, and W. Carrara, 2017. “Unleashing the potential of artificial intelligence in the public sector,” Capgemini Consulting, at https://www.capgemini.com/consulting/wp-content/uploads/sites/30/2017/10/ai-in-public-sector.pdf, accessed 26 December 2020.

J. Tito, 2017. “Destination unknown: Exploring the impact of artificial intelligence on government,” Centre for Public Impact, BGC Foundation (21 September), at https://www.centreforpublicimpact.org/ai-government-working-paper/, accessed 26 December 2020.

D. Valle-Cruz, 2019. “Public value of e-government services through emerging technologies,” International Journal of Public Sector Management, volume 32, number 5, pp. 530–545.
doi: https://doi.org/10.1108/IJPSM-03-2018-0072, accessed 26 December 2020.

D. Valle-Cruz, J.I. Criado, R. Sandoval-Almazán, and E.A. Ruvalcaba-Gomez, 2020. “Assessing the public policy-cycle framework in the age of artificial intelligence: From agenda-setting to policy evaluation,” Government Information Quarterly, volume 37, number 4, 101509.
doi: https://doi.org/10.1016/j.giq.2020.101509, accessed 26 December 2020.

D. Valle-Cruz, E. Ruvalcaba-Gomez, R. Sandoval-Almazán, and J.I. Criado, 2019. “A review of artificial intelligence in government and its potential from a public policy perspective,” dg.o 2019: Proceedings of the 20th Annual International Conference on Digital Government Research, pp. 91–99.
doi: https://doi.org/10.1145/3325112.3325242, accessed 26 December 2020.

S. Vydra and B. Klievink, 2019. “Techno-optimism and policy-pessimism in the public sector big data debate,” Government Information Quarterly, volume 36, number 4, 101383.
doi: https://doi.org/10.1016/j.giq.2019.05.010, accessed 26 December 2020.

P. Weill and S.L. Woerner, 2013. “The future of the CIO in a digital economy,” MIS Quarterly Executive, volume 12, number 2, article 3, at https://aisel.aisnet.org/misqe/vol12/iss2/3/, accessed 26 December 2020.

B. Williamson, 2014. “Knowing public services: Cross-sector intermediaries and algorithmic governance in public sector reform,” Public Policy and Administration, volume 29, number 4, pp. 292–312.
doi: https://doi.org/10.1177/0952076714529139, accessed 26 December 2020.

B.W. Wirtz, J.C. Weyerer, and C. Geyer, 2019. “Artificial intelligence and the public sector — Applications and challenges,” International Journal of Public Administration, volume 42, number 7, pp. 596–615.
doi: https://doi.org/10.1080/01900692.2018.1498103, accessed 26 December 2020.

N. Wright, 2018. “How artificial intelligence will reshape the global order,” Foreign Affairs (10 July), at https://www.foreignaffairs.com/articles/world/2018-07-10/how-artificial-intelligence-will-reshape-global-order, accessed 26 December 2020.

B. Zhang and A. Dafoe, 2019. “Artificial intelligence: American attitudes and trends,” SSRN (19 January).
doi: http://dx.doi.org/10.2139/ssrn.3312874, accessed 26 December 2020.

S. Zuboff, 2015. “Big other: Surveillance capitalism and the prospects of an information civilization,” Journal of Information Technology, volume 30, number 1, pp. 75–89.
doi: https://doi.org/10.1057/jit.2015.5, accessed 26 December 2020.

A. Zuurmond, 1994. “From bureaucracy to infocracy. A tale of two cities,” Informatization and the Public Sector volume 3, numbers 3–4, pp. 189–204.

 

Appendix

 

Dimensions of the analytical strategy and questions of the survey.
DimensionQuestionsItemsAnswers’ options
Perceptions about the concept of AI in the public sectorTo what extent do you agree with the following ideas regarding to artificial intelligence?— I am totally open to the adoption of AI.
— AI in the public sector is not very different than AI in the private sector.
— I assume naturally that robots and humans will share jobs.
— Envisioning future occupations of public employees is complex, as the replacement process will have profound impacts.
— Other “intelligences” will be required for public employees.
— The new professions in the public sector linked to AI will have to be tackled thoroughly.
1 being strongly disagree and 7 totally agree.
Regarding to this classification of human capacities and behaviors, to what extent do you agree with their replacement by artificial intelligence in the public sector?— Monitor
— Analyze
— Act
— Interact
— Remember
— Anticipate
— Feel
— Moralize
— Create
— Decide
1 being strongly disagree and 7 totally agree.
Expectations about AI in the public sectorFrom your point of view, in which public policy domains will be adopted artificial intelligence at an earliest stage?— Economic affairs
— Defense
— Education
— Mobility
— Security
— Citizens’ participation
— Culture promotion
— Environment protection
— Social policy
— Public health
— Tourism
— Housing and community services
Select at least three options.
From your point of view, what of the following functions will be more affected by artificial intelligence in the short term?— Clerical and assistant tasks
— Executive management
— Technical duties
— Training
— Management of organizational networks
— Public service delivery
— Regulation
— Processing of transactions/operations
Select at least three options.
Challenges and opportunities of AI in the public sectorWhat are the main inhibitors for the implementation of artificial intelligence in the public sector?— Digital divide
— Suppliers’ control
— Administrative culture
— Social unawareness
— Inequality
— Citizen literacy
— Human labor elimination
— Technological infrastructure
— Legislation
— Governance framework
— Budget
— Data privacy
— Data security
Select at least three options.
What are the main enablers for the implementation of artificial intelligence in the public sector?— Business competition
— Knowledge
— Administrative culture
— Citizen literacy
— Technological infrastructure
— Legislation
— Governance framework
— Budget
— Technological transfer
Select at least three options.

 

 


Editorial history

Received 1 July 2020; revised 27 July 2020; accepted 27 July 2020.


Creative Commons License
This paper is licensed under a Creative Commons Attribution 4.0 International License.

Chief information officers’ perceptions about artificial intelligence: A comparative study of implications and challenges for the public sector
by J. Ignacio Criado, Rodrigo Sandoval-Almazan, David Valle-Cruz, and Edgar A. Ruvalcaba-Gómez.
First Monday, Volume 26, Number 1 - 4 January 2021
https://journals.uic.edu/ojs/index.php/fm/article/download/10648/10042
doi: https://dx.doi.org/10.5210/fm.v26i1.10648