NIST Proposes Method for Evaluating User Trust in Artificial Intelligence Systems


Illustration shows how people evaluating two different tasks performed by AI -- music selection and medical diagnosis -- might trust the AI varying amounts because the risk level of each task is different.
NIST’s new publication proposes a list of nine factors that contribute to a human’s potential trust in an AI system. A person may weigh the nine factors differently depending on both the task itself and the risk involved in trusting the AI’s decision. As an example, two different AI programs — a music selection algorithm and an AI that assists with cancer diagnosis — may score the same on all nine criteria. Users, however, might be inclined to trust the music selection algorithm but not the medical assistant, which is performing a far riskier task.Credit: N. Hanacek/NIST

National Institute of Standards and Technology (NIST): ” Every time you speak to a virtual assistant on your smartphone, you are talking to an artificial intelligence — an AI that can, for example, learn your taste in music and make song recommendations that improve based on your interactions. However, AI also assists us with more risk-fraught activities, such as helping doctors diagnose cancer. These are two very different scenarios, but the same issue permeates both: How do we humans decide whether or not to trust a machine’s recommendations? 

This is the question that a new draft publication from the National Institute of Standards and Technology (NIST) poses, with the goal of stimulating a discussion about how humans trust AI systems. The document, Artificial Intelligence and User Trust (NISTIR 8332), is open for public comment until July 30, 2021. 

The report contributes to the broader NIST effort to help advance trustworthy AI systems. The focus of this latest publication is to understand how humans experience trust as they use or are affected by AI systems….(More)”.

Experts Doubt Ethical AI Design Will Be Broadly Adopted as the Norm Within the Next Decade


Report by Pew Research Center: “Artificial intelligence systems “understand” and shape a lot of what happens in people’s lives. AI applications “speak” to people and answer questions when the name of a digital voice assistant is called out. They run the chatbots that handle customer-service issues people have with companies. They help diagnose cancer and other medical conditions. They scour the use of credit cards for signs of fraud, and they determine who could be a credit risk.

They help people drive from point A to point B and update traffic information to shorten travel times. They are the operating system of driverless vehicles. They sift applications to make recommendations about job candidates. They determine the material that is offered up in people’s newsfeeds and video choices.

They recognize people’s facestranslate languages and suggest how to complete people’s sentences or search queries. They can “read” people’s emotions. They beat them at sophisticated games. They write news stories, paint in the style of Vincent Van Gogh and create music that sounds quite like the Beatles and Bach.

Corporations and governments are charging evermore expansively into AI development. Increasingly, nonprogrammers can set up off-the-shelf, pre-built AI tools as they prefer.

As this unfolds, a number of experts and advocates around the world have become worried about the long-term impact and implications of AI applications. They have concerns about how advances in AI will affect what it means to be human, to be productive and to exercise free will. Dozens of convenings and study groups have issued papers proposing what the tenets of ethical AI design should be, and government working teams have tried to address these issues. In light of this, Pew Research Center and Elon University’s Imagining the Internet Center asked experts where they thought efforts aimed at creating ethical artificial intelligence would stand in the year 2030….(More)”

Citizens ‘on mute’ in digital public service delivery


Blog by Sarah Giest at Data and Policy: “Various countries are digitalizing their welfare system in the larger context of austerity considerations and fraud detection goals, but these changes are increasingly under scrutiny. In short, digitalization of the welfare system means that with the help of mathematical models, data and/or the combination of different administrative datasets, algorithms issue a decision on, for example, an application for social benefits (Dencik and Kaun 2020).

Several examples exist where such systems have led to unfair treatment of welfare recipients. In Europe, the Dutch SyRI system has been banned by court, due to human rights violations in the profiling of welfare recipients, and the UK has found errors in the automated processes leading to financial hardship among citizens. In the United States and Canada, automated systems led to false underpayment or denial of benefits. A recent UN report (2019) even warns that countries are ‘stumbling zombie-like into a digital welfare dystopia’. Further, studies raise alarm that this process of digitalization is done in a way that it not only creates excessive information asymmetry among government and citizens, but also disadvantages certain groups more than others.

A closer look at the Dutch Childcare Allowance case highlights this. In this example, low-income parents were regarded as fraudsters by the Tax Authorities if they had incorrectly filled out any documents. An automated and algorithm-based procedure then also singled out dual-nationality families. The victims lost their allowance without having been given any reasons. Even worse, benefits already received were reclaimed. This led to individual hardship, where financial troubles and the categorization as a fraudster by government led for citizens to a chain of events from unpaid healthcare insurance and the inability to visit a doctor to job loss, potential home loss and mental health concerns (Volkskrant 2020)….(More)”.

Selected Readings on the Use of Artificial Intelligence in the Public Sector


By Kateryna Gazaryan and Uma Kalkar

The Living Library’s Selected Readings series seeks to build a knowledge base on innovative approaches for improving the effectiveness and legitimacy of governance. This curated and annotated collection of recommended works focuses on algorithms and artificial intelligence in the public sector.

As Artificial Intelligence becomes more developed, governments have turned to it to improve the speed and quality of public sector service delivery, among other objectives. Below, we provide a selection of recent literature that examines how the public sector has adopted AI to serve constituents and solve public problems. While the use of AI in governments can cut down costs and administrative work, these technologies are often early in development and difficult for organizations to understand and control with potential harmful effects as a result. As such, this selected reading explores not only the use of artificial intelligence in governance but also its benefits, and its consequences.

Readings are listed in alphabetical order.

Berryhill, Jamie, Kévin Kok Heang, Rob Clogher, and Keegan McBride. “Hello, World: Artificial intelligence and its use in the public sector.OECD Working Papers on Public Governance no. 36 (2019): https://doi.org/10.1787/726fd39d-en.

This working paper emphasizes the importance of defining AI for the public sector and outlining use cases of AI within governments. It provides a map of 50 countries that have implemented or set in motion the development of AI strategies and highlights where and how these initiatives are cross-cutting, innovative, and dynamic. Additionally, the piece provides policy recommendations governments should consider when exploring public AI strategies to adopt holistic and humanistic approaches.

Kuziemski, Maciej, and Gianluca Misuraca. “AI Governance in the Public Sector: Three Tales from the Frontiers of Automated Decision-Making in Democratic Settings.” Telecommunications Policy 44, no. 6 (2020): 101976. 

Kuziemski and Misuraca explore how the use of artificial intelligence in the public sector can exacerbate existing power imbalances between the public and the government. They consider the European Union’s artificial intelligence “governance and regulatory frameworks” and compare these policies with those of Canada, Finland, and Poland. Drawing on previous scholarship, the authors outline the goals, drivers, barriers, and risks of incorporating artificial intelligence into public services and assess existing regulations against these factors. Ultimately, they find that the “current AI policy debate is heavily skewed towards voluntary standards and self-governance” while minimizing the influence of power dynamics between governments and constituents. 

Misuraca, Gianluca, and Colin van Noordt. “AI Watch, Artificial Intelligence in Public Services: Overview of the Use and Impact of AI in Public Services in the EU.” 30255 (2020).

This study provides “evidence-based scientific support” for the European Commission as it navigates AI regulation via an overview of ways in which European Union member-states use AI to enhance their public sector operations. While AI has the potential to positively disrupt existing policies and functionalities, this report finds gaps in how AI gets applied by governments. It suggests the need for further research centered on the humanistic, ethical, and social ramification of AI use and a rigorous risk assessment from a “public-value perspective” when implementing AI technologies. Additionally, efforts must be made to empower all European countries to adopt responsible and coherent AI policies and techniques.

Saldanha, Douglas Morgan Fullin, and Marcela Barbosa da Silva. “Transparency and Accountability of Government Algorithms: The Case of the Brazilian Electronic Voting System.” Cadernos EBAPE.BR 18 (2020): 697–712.

Saldanha and da Silva note that open data and open government revolutions have increased citizen demand for algorithmic transparency. Algorithms are increasingly used by governments to speed up processes and reduce costs, but their black-box  systems and lack of explanability allows them to insert implicit and explicit bias and discrimination into their calculations. The authors conduct a qualitative study of the “practices and characteristics of the transparency and accountability” in the Brazilian e-voting system across seven dimensions: consciousness; access and reparations; accountability; explanation; data origin, privacy and justice; auditing; and validation, precision and tests. They find the Brazilian e-voting system fulfilled the need to inform citizens about the benefits and consequences of data collection and algorithm use but severely lacked in demonstrating accountability and opening algorithm processes for citizen oversight. They put forth policy recommendations to increase the e-voting system’s accountability to Brazilians and strengthen auditing and oversight processes to reduce the current distrust in the system.

Sharma, Gagan Deep, Anshita Yadav, and Ritika Chopra. “Artificial intelligence and effective governance: A review, critique and research agenda.Sustainable Futures 2 (2020): 100004.

This paper conducts a systematic review of the literature of how AI is used across different branches of government, specifically, healthcare, information, communication, and technology, environment, transportation, policy making, and economic sectors. Across the 74 papers surveyed, the authors find a gap in the research on selecting and implementing AI technologies, as well as their monitoring and evaluation. They call on future research to assess the impact of AI pre- and post-adoption in governance, along with the risks and challenges associated with the technology.

Tallerås, Kim, Terje Colbjørnsen, Knut Oterholm, and Håkon Larsen. “Cultural Policies, Social Missions, Algorithms and Discretion: What Should Public Service Institutions Recommend?Part of the Lecture Notes in Computer Science book series (2020).

Tallerås et al. examine how the use of algorithms by public services, such as public radio and libraries, influence broader society and culture. For instance, to modernize their offerings, Norway’s broadcasting corporation (NRK) has adopted online platforms similar to popular private streaming services. However, NRK’s filtering process has faced “exposure diversity” problems that narrow recommendations to already popular entertainment and move Norway’s cultural offerings towards a singularity. As a public institution, NRK is required to “fulfill […] some cultural policy goals,” raising the question of how public media services can remain relevant in the era of algorithms fed by “individualized digital culture.” Efforts are currently underway to employ recommendation systems that balance cultural diversity with personalized content relevance that engage individuals and uphold the socio-cultural mission of public media.

Vogl, Thomas, Seidelin Cathrine, Bharath Ganesh, and Jonathan Bright. “Smart Technology and the Emergence of Algorithmic Bureaucracy: Artificial Intelligence in UK Local Authorities.” Public administration review 80, no. 6 (2020): 946–961.

Local governments are using “smart technologies” to create more efficient and effective public service delivery. These tools are twofold: not only do they help the public interact with local authorities, they also streamline the tasks of government officials. To better understand the digitization of local government, the authors conducted surveys, desk research, and in-depth interviews with stakeholders from local British governments to understand reasoning, processes, and experiences within a changing government framework. Vogl et al. found an increase in “algorithmic bureaucracy” at the local level to reduce administrative tasks for government employees, generate feedback loops, and use data to enhance services. While the shift toward digital local government demonstrates initiatives to utilize emerging technology for public good, further research is required to determine which demographics are not involved in the design and implementation of smart technology services and how to identify and include these audiences.

Wirtz, Bernd W., Jan C. Weyerer, and Carolin Geyer. “Artificial intelligence and the public sector—Applications and challenges.International Journal of Public Administration 42, no. 7 (2019): 596-615.

The authors provide an extensive review of the existing literature on AI uses and challenges in the public sector to identify the gaps in current applications. The developing nature of AI in public service has led to differing definitions of what constitutes AI and what are the risks and benefits it poses to the public. As well, the authors note the lack of focus on the downfalls of AI in governance, with studies tending to primarily focus on the positive aspects of the technology. From this qualitative analysis, the researchers highlight ten AI applications: knowledge management, process automation, virtual agents, predictive analytics and data visualization, identity analytics, autonomous systems, recommendation systems, digital assistants, speech analytics, and threat intelligence. As well, they note four challenge dimensions—technology implementation, laws and regulation, ethics, and society. From these applications and risks, Wirtz et al. provide a “checklist for public managers” to make informed decisions on how to integrate AI into their operations. 

Wirtz, Bernd W., Jan C. Weyerer, and Benjamin J. Sturm. “The dark sides of artificial intelligence: An integrated AI governance framework for public administration.International Journal of Public Administration 43, no. 9 (2020): 818-829.

As AI is increasingly popularized and picked up by governments, Wirtz et al. highlight the lack of research on the challenges and risks—specifically, privacy and security—associated with implementing AI systems in the public sector. After assessing existing literature and uncovering gaps in the main governance frameworks, the authors outline the three areas of challenges of public AI: law and regulations, society, and ethics. Last, they propose an “integrated AI governance framework” that takes into account the risks of AI for a more holistic “big picture” approach to AI in the public sector.

Zuiderwijk, Anneke, Yu-Che Chen, and Fadi Salem. “Implications of the use of artificial intelligence in public governance: A systematic literature review and a research agenda.Government Information Quarterly (2021): 101577.

Following a literature review on the risks and possibilities of AI in the public sector, Zuiderwijk, Chen, and Salem design a research agenda centered around the “implications of the use of AI for public governance.” The authors provide eight process recommendations, including: avoiding superficial buzzwords in research; conducting domain- and locality-specific research on AI in governance; shifting from qualitative analysis to diverse research methods; applying private sector “practice-driven research” to public sector study; furthering quantitative research on AI use by governments; creating “explanatory research designs”; sharing data for broader study; and adopting multidisciplinary reference theories. Further, they note the need for scholarship to delve into best practices, risk management, stakeholder communication, multisector use, and impact assessments of AI in the public sector to help decision-makers make informed decisions on the introduction, implementation, and oversight of AI in the public sector.

Introducing the AI Localism Repository


The GovLab: “Artificial intelligence is here to stay. As this technology advances—both in its complexity and ubiquity across our societies—decision-makers must address the growing nuances of AI regulation and oversight. Early last year, The GovLab’s Stefaan Verhulst and Mona Sloane coined the term “AI localism” to describe how local governments have stepped up to regulate AI policies, design governance frameworks, and monitor AI use in the public sector. 

While top-level regulation remains scant, many municipalities have taken to addressing AI use in their communities. Today, The GovLab is proud to announce the soft launch of the AI Localism Repository. This living platform is a curated collection of AI localism initiatives across the globe categorized by geographic regions, types of technological and governmental innovation in AI regulation, mechanisms of governance, and sector focus. 

We invite visitors to explore this repository and learn more about the inventive measures cities are taking to control how, when, and why AI is being used by public authorities. We also welcome additional case study submissions, which can be sent to us via Google Form….(More)”

The Constitution of Algorithms


Open Access Book by By Florian Jaton: “A laboratory study that investigates how algorithms come into existence. Algorithms—often associated with the terms big datamachine learning, or artificial intelligence—underlie the technologies we use every day, and disputes over the consequences, actual or potential, of new algorithms arise regularly. In this book, Florian Jaton offers a new way to study computerized methods, providing an account of where algorithms come from and how they are constituted, investigating the practical activities by which algorithms are progressively assembled rather than what they may suggest or require once they are assembled.

Drawing on a four-year ethnographic study of a computer science laboratory that specialized in digital image processing, Jaton illuminates the invisible processes that are behind the development of algorithms. Tracing what he terms a set of intertwining courses of actions sharing common finalities, he describes the practical activity of creating algorithms through the lenses of ground-truthingprogramming, and formulating. He first presents the building of ground truths, referential repositories that form the material basis for algorithms. Then, after considering programming’s resistance to ethnographic scrutiny, he describes programming courses of action he attended at the laboratory. Finally, he offers an account of courses of action that successfully formulated some of the relationships among the data of a ground-truth database, revealing the links between ground-truthing, programming, and formulating activities—entangled processes that lead to the shaping of algorithms. In practice, ground-truthing, programming, and formulating form a whirlwind process, an emergent and intertwined agency….(More)”.

AI and Shared Prosperity


Paper by Katya Klinova and Anton Korinek: “Future advances in AI that automate away human labor may have stark implications for labor markets and inequality. This paper proposes a framework to analyze the effects of specific types of AI systems on the labor market, based on how much labor demand they will create versus displace, while taking into account that productivity gains also make society wealthier and thereby contribute to additional labor demand. This analysis enables ethically-minded companies creating or deploying AI systems as well as researchers and policymakers to take into account the effects of their actions on labor markets and inequality, and therefore to steer progress in AI in a direction that advances shared prosperity and an inclusive economic future for all of humanity…(More)”.

AI helps scour video archives for evidence of human-rights abuses


The Economist: “Thanks especially to ubiquitous camera-phones, today’s wars have been filmed more than any in history. Consider the growing archives of Mnemonic, a Berlin charity that preserves video that purports to document war crimes and other violations of human rights. If played nonstop, Mnemonic’s collection of video from Syria’s decade-long war would run until 2061. Mnemonic also holds seemingly bottomless archives of video from conflicts in Sudan and Yemen. Even greater amounts of potentially relevant additional footage await review online.

Outfits that, like Mnemonic, scan video for evidence of rights abuses note that the task is a slog. Some trim costs by recruiting volunteer reviewers. Not everyone, however, is cut out for the tedium and, especially, periodic dreadfulness involved. That is true even for paid staff. Karim Khan, who leads a United Nations team in Baghdad investigating Islamic State (IS) atrocities, says viewing the graphic cruelty causes enough “secondary trauma” for turnover to be high. The UN project, called UNITAD, is sifting through documentation that includes more than a year’s worth of video, most of it found online or on the phones and computers of captured or killed IS members.

Now, however, reviewing such video is becoming much easier. Technologists are developing a type of artificial-intelligence (AI) software that uses “machine vision” to rapidly scour video for imagery that suggests an abuse of human rights has been recorded. It’s early days, but the software is promising. A number of organisations, including Mnemonic and UNITAD, have begun to operate such programs.

This year UNITAD began to run one dubbed Zeteo. It performs well, says David Hasman, one of its operators. Zeteo can be instructed to find—and, if the image resolution is decent, typically does find—bits of video showing things like explosions, beheadings, firing into a crowd and grave-digging. Zeteo can also spot footage of a known person’s face, as well as scenes as precise as a woman walking in uniform, a boy holding a gun in twilight, and people sitting on a rug with an IS flag in view. Searches can encompass metadata that reveals when, where and on what devices clips were filmed….(More)”.

Confronting Bias: BSA’s Framework to Build Trust in AI


BSA Software Alliance: “The Framework is a playbook organizations can use to enhance trust in their AI systems through risk management processes that promote fairness, transparency, and accountability. It can be leveraged by organizations that develop AI systems and companies that acquire and deploy such systems as the basis for:
– Internal Process Guidance. The Framework can be used as a tool for organizing and establishing roles,
responsibilities, and expectations for internal risk management processes.
– Training, Awareness, and Education. The Framework can be used to build internal training and education
programs for employees involved in developing and using AI systems, and for educating executives about
the organization’s approach to managing AI bias risks.
– Supply Chain Assurance and Accountability. AI developers and organizations that deploy AI
systems can use the Framework as a basis for communicating and coordinating about their respective roles and responsibilities for managing AI risks throughout a system’s lifecycle.
– Trust and Confidence. The Framework can help organizations communicate information about a
product’s features and its approach to mitigating AI bias risks to a public audience. In that sense, the
Framework can help organizations communicate to the public about their commitment to building
ethical AI systems.
– Incident Response. Following an unexpected incident, the processes and documentation set forth
in the Framework can serve as an audit trail that can help organizations quickly diagnose and remediate
potential problems…(More)”

Implications of the use of artificial intelligence in public governance: A systematic literature review and a research agenda


Paper by Anneke Zuiderwijk, Yu-Che Chen and Fadi Salem: “To lay the foundation for the special issue that this research article introduces, we present 1) a systematic review of existing literature on the implications of the use of Artificial Intelligence (AI) in public governance and 2) develop a research agenda. First, an assessment based on 26 articles on this topic reveals much exploratory, conceptual, qualitative, and practice-driven research in studies reflecting the increasing complexities of using AI in government – and the resulting implications, opportunities, and risks thereof for public governance. Second, based on both the literature review and the analysis of articles included in this special issue, we propose a research agenda comprising eight process-related recommendations and seven content-related recommendations. Process-wise, future research on the implications of the use of AI for public governance should move towards more public sector-focused, empirical, multidisciplinary, and explanatory research while focusing more on specific forms of AI rather than AI in general. Content-wise, our research agenda calls for the development of solid, multidisciplinary, theoretical foundations for the use of AI for public governance, as well as investigations of effective implementation, engagement, and communication plans for government strategies on AI use in the public sector. Finally, the research agenda calls for research into managing the risks of AI use in the public sector, governance modes possible for AI use in the public sector, performance and impact measurement of AI use in government, and impact evaluation of scaling-up AI usage in the public sector….(More)”.