Paper by Davy Tsz KitNg, Jac Ka LokLeung, Samuel K.W.Chu, and Maggie QiaoShen: “Artificial Intelligence (AI) has spread across industries (e.g., business, science, art, education) to enhance user experience, improve work efficiency, and create many future job opportunities. However, public understanding of AI technologies and how to define AI literacy is under-explored. This vision poses upcoming challenges for our next generation to learn about AI. On this note, an exploratory review was conducted to conceptualize the newly emerging concept “AI literacy”, in search for a sound theoretical foundation to define, teach and evaluate AI literacy. Grounded in literature on 30 existing peer-reviewed articles, this review proposed four aspects (i.e., know and understand, use, evaluate, and ethical issues) for fostering AI literacy based on the adaptation of classic literacies. This study sheds light on the consolidated definition, teaching, and ethical concerns on AI literacy, establishing the groundwork for future research such as competency development and assessment criteria on AI literacy….(More)”.
Automating Decision-making in Migration Policy: A Navigation Guide
Report by Astrid Ziebarth and Jessica Bither: “Algorithmic-driven or automated decision-making models (ADM) and programs are increasingly used by public administrations to assist human decision-making processes in public policy—including migration and refugee policy. These systems are often presented as a neutral, technological fix to make policy and systems more efficient. However, migration policymakers and stakeholders often do not understand exactly how these systems operate. As a result, the implications of adopting ADM technology are still unclear, and sometimes not considered. In fact, automated decision-making systems are never neutral, nor is their employment inevitable. To make sense of their function and decide whether or how to use them in migration policy will require consideration of the specific context in which ADM systems are being employed.
Three concrete use cases at core nodes of migration policy in which automated decision-making is already either being developed or tested are examined: visa application processes, placement matching to improve integration outcomes, and forecasting models to assist for planning and preparedness related to human mobility or displacement. All cases raise the same categories of questions: from the data employed, to the motivation behind using a given system, to the action triggered by models. The nuances of each case demonstrate why it is crucial to understand these systems within a bigger socio-technological context and provide categories and questions that can help policymakers understand the most important implications of any new system, including both technical consideration (related to accuracy, data questions, or bias) as well as contextual questions (what are we optimizing for?).
Stakeholders working in the migration and refugee policy space must make more direct links to current discussions surrounding governance, regulation of AI, and digital rights more broadly. We suggest some first points of entry toward this goal. Specifically, for next steps stakeholders should:
- Bridge migration policy with developments in digital rights and tech regulation
- Adapt emerging policy tools on ADM to migration space
- Create new spaces for exchange between migration policymakers, tech regulators, technologists, and civil society
- Include discussion on the use of ADM systems in international migration fora
- Increase the number of technologists or bilinguals working in migration policy
- Link tech and migration policy to bigger questions of foreign policy and geopolitics…(More)”.
New York City passed a bill requiring ‘bias audits’ of AI hiring tech
Kate Kaye at Protocol: “Let the AI auditing vendor brigade begin. A year since it was introduced, New York City Council passed a bill earlier this week requiring companies that sell AI technologies for hiring to obtain audits assessing the potential of those products to discriminate against job candidates. The bill requiring “bias audits” passed with overwhelming support in a 38-4 vote.
The bill is intended to weed out the use of tools that enable already unlawful employment discrimination in New York City. If signed into law, it will require providers of automated employment decision tools to have those systems evaluated each year by an audit service and provide the results to companies using those systems.
AI for recruitment can include software that uses machine learning to sift through resumes and help make hiring decisions, systems that attempt to decipher the sentiments of a job candidate, or even tech involving games to pick up on subtle clues about someone’s hiring worthiness. The NYC bill attempts to encompass the full gamut of AI by covering everything from old-school decision trees to more complex systems operating through neural networks.
The legislation calls on companies using automated decision tools for recruitment not only to tell job candidates when they’re being used, but to tell them what information the technology used to evaluate their suitability for a job.
The bill, however, fails to go into detail on what constitutes a bias audit other than to define one as “an impartial evaluation” that involves testing. And it already has critics who say it was rushed into passage and doesn’t address discrimination related to disability or age…(More)”.
AI-tocracy
Paper by Martin Beraja, Andrew Kao, David Y. Yang & Noam Yuchtman: “Can frontier innovation be sustained under autocracy? We argue that innovation and autocracy can be mutually reinforcing when: (i) the new technology bolsters the autocrat’s power; and (ii) the autocrat’s demand for the technology stimulates further innovation in applications beyond those benefiting it directly. We test for such a mutually reinforcing relationship in the context of facial recognition AI in China. To do so, we gather comprehensive data on AI firms and government procurement contracts, as well as on social unrest across China during the last decade. We first show that autocrats benefit from AI: local unrest leads to greater government procurement of facial recognition AI, and increased AI procurement suppresses subsequent unrest. We then show that AI innovation benefits from autocrats’ suppression of unrest: the contracted AI firms innovate more both for the government and commercial markets. Taken together, these results suggest the possibility of sustained AI innovation under the Chinese regime: AI innovation entrenches the regime, and the regime’s investment in AI for political control stimulates further frontier innovation….(More)”.
‘Is it OK to …’: the bot that gives you an instant moral judgment
Article by Poppy Noor: “Corporal punishment, wearing fur, pineapple on pizza – moral dilemmas, are by their very nature, hard to solve. That’s why the same ethical questions are constantly resurfaced in TV, films and literature.
But what if AI could take away the brain work and answer ethical quandaries for us? Ask Delphi is a bot that’s been fed more than 1.7m examples of people’s ethical judgments on everyday questions and scenarios. If you pose an ethical quandary, it will tell you whether something is right, wrong, or indefensible.
Anyone can use Delphi. Users just put a question to the bot on its website, and see what it comes up with.
The AI is fed a vast number of scenarios – including ones from the popular Am I The Asshole sub-Reddit, where Reddit users post dilemmas from their personal lives and get an audience to judge who the asshole in the situation was.
Then, people are recruited from Mechanical Turk – a market place where researchers find paid participants for studies – to say whether they agree with the AI’s answers. Each answer is put to three arbiters, with the majority or average conclusion used to decide right from wrong. The process is selective – participants have to score well on a test to qualify to be a moral arbiter, and the researchers don’t recruit people who show signs of racism or sexism.
The arbitrators agree with the bot’s ethical judgments 92% of the time (although that could say as much about their ethics as it does the bot’s)…(More)”.
AI Generates Hypotheses Human Scientists Have Not Thought Of
Robin Blades in Scientific American: “Electric vehicles have the potential to substantially reduce carbon emissions, but car companies are running out of materials to make batteries. One crucial component, nickel, is projected to cause supply shortages as early as the end of this year. Scientists recently discovered four new materials that could potentially help—and what may be even more intriguing is how they found these materials: the researchers relied on artificial intelligence to pick out useful chemicals from a list of more than 300 options. And they are not the only humans turning to A.I. for scientific inspiration.
Creating hypotheses has long been a purely human domain. Now, though, scientists are beginning to ask machine learning to produce original insights. They are designing neural networks (a type of machine-learning setup with a structure inspired by the human brain) that suggest new hypotheses based on patterns the networks find in data instead of relying on human assumptions. Many fields may soon turn to the muse of machine learning in an attempt to speed up the scientific process and reduce human biases.
In the case of new battery materials, scientists pursuing such tasks have typically relied on database search tools, modeling and their own intuition about chemicals to pick out useful compounds. Instead a team at the University of Liverpool in England used machine learning to streamline the creative process. The researchers developed a neural network that ranked chemical combinations by how likely they were to result in a useful new material. Then the scientists used these rankings to guide their experiments in the laboratory. They identified four promising candidates for battery materials without having to test everything on their list, saving them months of trial and error…(More)”.
Climate Change and AI: Recommendations for Government
Press Release: “A new report, developed by the Centre for AI & Climate and Climate Change AI for the Global Partnership on AI (GPAI), calls for governments to recognise the potential for artificial intelligence (AI) to accelerate the transition to net zero, and to put in place the support needed to advance AI-for-climate solutions. The report is being presented at COP26 today.
The report, Climate Change and AI: Recommendations for Government, highlights 48 specific recommendations for how governments can both support the application of AI to climate challenges and address the climate-related risks that AI poses.
The report was commissioned by the Global Partnership on AI (GPAI), a partnership between 18 countries and the EU that brings together experts from across countries and sectors to help shape the development of AI.
AI is already being used to support climate action in a wide range of use cases, several of which the report highlights. These include:
- National Grid ESO, which has used AI to double the accuracy of its forecasts of UK electricity demand. Radically improving forecasts of electricity demand and renewable energy generation will be critical in enabling greater proportions of renewable energy on electricity grids.
- The UN Satellite Centre (UNOSAT), which has developed the FloodAI system that delivers high-frequency flood reports. FloodAI’s reports, which use a combination of satellite data and machine learning, have improved the response to climate-related disasters in Asia and Africa.
- Climate TRACE, a global coalition of organizations, which has radically improved the transparency and accuracy of emissions monitoring by leveraging AI algorithms and data from more than 300 satellites and 11,000 sensors.
The authors also detail critical bottlenecks that are impeding faster adoption. To address these, the report calls for governments to:
- Improve data ecosystems in sectors critical to climate transition, including the development of digital twins in e.g. the energy sector.
- Increase support for research, innovation, and deployment through targeted funding, infrastructure, and improved market designs.
- Make climate change a central consideration in AI strategies to shape the responsible development of AI as a whole.
- Support greater international collaboration and capacity building to facilitate the development and governance of AI-for-climate solutions….(More)”.
Why Are We Failing at AI Ethics?
Article by Anja Kaspersen and Wendell Wallach: “…Extremely troubling is the fact that the people who are most vulnerable to negative impacts from such rapid expansions of AI systems are often the least likely to be able to join the conversation about these systems, either because they have no or restricted digital access or their lack of digital literacy makes them ripe for exploitation.
Such vulnerable groups are often theoretically included in discussions, but not empowered to take a meaningful part in making decisions. This engineered inequity, alongside human biases, risks amplifying otherness through neglect, exclusion, misinformation, and disinformation.
Society should be deeply concerned that nowhere near enough substantive progress is being made to develop and scale actionable legal, ethical oversight while simultaneously addressing existing inequalities.
So, why hasn’t more been done? There are three main issues at play:
First, many of the existing dialogues around the ethics of AI and governance are too narrow and fail to understand the subtleties and life cycles of AI systems and their impacts.
Often, these efforts focus only on the development and deployment stages of the technology life cycle, when many of the problems occur during the earlier stages of conceptualization, research, and design. Or they fail to comprehend when and if AI system operates at a level of maturity required to avoid failure in complex adaptive systems.
Or they focus on some aspects of ethics, while ignoring other aspects that are more fundamental and challenging. This is the problem known as “ethics washing” – creating a superficially reassuring but illusory sense that ethical issues are being adequately addressed, to justify pressing forward with systems that end up deepening current patterns.
Let’s be clear: every choice entails tradeoffs. “Ethics talk” is often about underscoring the various tradeoffs entailed in differing courses of action. Once a course has been selected, comprehensive ethical oversight is also about addressing the considerations not accommodated by the options selected, which is essential to any future verification effort. This vital part of the process is often a stumbling block for those trying to address the ethics of AI.
The second major issue is that to date all the talk about ethics is simply that: talk.
We’ve yet to see these discussions translate into meaningful change in managing the ways in which AI systems are being embedded into various aspect of our lives….
A third issue at play is that discussions on AI and ethics are still largely confined to the ivory tower.
There is an urgent need for more informed public discourse and serious investment in civic education around the societal impact of the bio-digital revolution. This could help address the first two problems, but most of what the general public currently perceives about AI comes from sci-fi tropes and blockbuster movies.
A few examples of algorithmic bias have penetrated the public discourse. But the most headline-grabbing research on AI and ethics tends to focus on far-horizon existential risks. More effort needs to be invested in communicating to the public that, beyond the hypothetical risks of future AI, there are real and imminent risks posed by why and how we embed AI systems that currently shape everyone’s daily lives….(More)”.
Conceptual and normative approaches to AI governance for a global digital ecosystem supportive of the UN Sustainable Development Goals (SDGs)
Paper by Amandeep S. Gill & Stefan Germann: “AI governance is like one of those mythical creatures that everyone speaks of but which no one has seen. Sometimes, it is reduced to a list of shared principles such as transparency, non-discrimination, and sustainability; at other times, it is conflated with specific mechanisms for certification of algorithmic solutions or ways to protect the privacy of personal data. We suggest a conceptual and normative approach to AI governance in the context of a global digital public goods ecosystem to enable progress on the UN Sustainable Development Goals (SDGs). Conceptually, we propose rooting this approach in the human capability concept—what people are able to do and to be, and in a layered governance framework connecting the local to the global. Normatively, we suggest the following six irreducibles: a. human rights first; b. multi-stakeholder smart regulation; c. privacy and protection of personal data; d. a holistic approach to data use captured by the 3Ms—misuse of data, missed use of data and missing data; e. global collaboration (‘digital cooperation’); f. basing governance more in practice, in particular, thinking separately and together about data and algorithms. Throughout the article, we use examples from the health domain particularly in the current context of the Covid-19 pandemic. We conclude by arguing that taking a distributed but coordinated global digital commons approach to the governance of AI is the best guarantee of citizen-centered and societally beneficial use of digital technologies for the SDGs…(More)”.
Understanding Algorithmic Discrimination in Health Economics Through the Lens of Measurement Errors
Paper by Anirban Basu, Noah Hammarlund, Sara Khor & Aasthaa Bansal: “There is growing concern that the increasing use of machine learning and artificial intelligence-based systems may exacerbate health disparities through discrimination. We provide a hierarchical definition of discrimination consisting of algorithmic discrimination arising from predictive scores used for allocating resources and human discrimination arising from allocating resources by human decision-makers conditional on these predictive scores. We then offer an overarching statistical framework of algorithmic discrimination through the lens of measurement errors, which is familiar to the health economics audience. Specifically, we show that algorithmic discrimination exists when measurement errors exist in either the outcome or the predictors, and there is endogenous selection for participation in the observed data. The absence of any of these phenomena would eliminate algorithmic discrimination. We show that although equalized odds constraints can be employed as bias-mitigating strategies, such constraints may increase algorithmic discrimination when there is measurement error in the dependent variable….(More)”.