Machine learning models are used in many sensitive areas where besides predictive accuracy their comprehensibility is also important. Interpretability of prediction models is necessary to determine their biases and causes of errors, and is a necessary prerequisite for users' confidence. For complex state-of-the-art black-box models post-hoc model-independent explanation techniques are an established solution. Popular and effective techniques, such as IME, LIME, and SHAP, use perturbation of instance features to explain individual predictions. Recently, Slack et al. (2020) put their robustness into question by showing that their outcomes can be manipulated due to poor perturbation sampling employed. This weakness would allow dieselgate type cheating of owners of sensitive models who could deceive inspection and hide potentially unethical or illegal biases existing in their predictive models. This could undermine public trust in machine learning models and give rise to legal restrictions on their use. We show that better sampling in these explanation methods prevents malicious manipulations. The proposed sampling uses data generators that learn the training set distribution and generate new perturbation instances much more similar to the training set. We show that the improved sampling increases the robustness of the LIME and SHAP, while previously untested method IME is already the most robust of all.

Conventional algorithmic fairness is West-centric, as seen in its sub-groups, values, and methods. In this paper, we de-center algorithmic fairness and analyse AI power in India. Based on 36 qualitative interviews and a discourse analysis of algorithmic deployments in India, we find that several assumptions of algorithmic fairness are challenged. We find that in India, data is not always reliable due to socio-economic factors, ML makers appear to follow double standards, and AI evokes unquestioning aspiration. We contend that localising model fairness alone can be window dressing in India, where the distance between models and oppressed communities is large. Instead, we re-imagine algorithmic fairness in India and provide a roadmap to re-contextualise data and models, empower oppressed communities, and enable Fair-ML ecosystems.

Disagreements are pervasive in human communication. In this paper we investigate what makes disagreement constructive. To this end, we construct WikiDisputes, a corpus of 7 425 Wikipedia Talk page conversations that contain content disputes, and define the task of predicting whether disagreements will be escalated to mediation by a moderator. We evaluate feature-based models with linguistic markers from previous work, and demonstrate that their performance is improved by using features that capture changes in linguistic markers throughout the conversations, as opposed to averaged values. We develop a variety of neural models and show that taking into account the structure of the conversation improves predictive accuracy, exceeding that of feature-based models. We assess our best neural model in terms of both predictive accuracy and uncertainty by evaluating its behaviour when it is only exposed to the beginning of the conversation, finding that model accuracy improves and uncertainty reduces as models are exposed to more information.

When engaging in argumentative discourse, skilled human debaters tailor claims to the beliefs of the audience, to construct effective arguments. Recently, the field of computational argumentation witnessed extensive effort to address the automatic generation of arguments. However, existing approaches do not perform any audience-specific adaptation. In this work, we aim to bridge this gap by studying the task of belief-based claim generation: Given a controversial topic and a set of beliefs, generate an argumentative claim tailored to the beliefs. To tackle this task, we model the people's prior beliefs through their stances on controversial topics and extend state-of-the-art text generation models to generate claims conditioned on the beliefs. Our automatic evaluation confirms the ability of our approach to adapt claims to a set of given beliefs. In a manual study, we additionally evaluate the generated claims in terms of informativeness and their likelihood to be uttered by someone with a respective belief. Our results reveal the limitations of modeling users' beliefs based on their stances, but demonstrate the potential of encoding beliefs into argumentative texts, laying the ground for future exploration of audience reach.

Various Alphabet unions have joined forces and created a global alliance called Alpha Global. It was created with some help from UNI Global, a federation of unions that represents 20 million workers from a variety of fields including the technology industry. According to The Verge, 13 different unions are part of Alpha Global, spanning 10 different countries including the US, UK, Germany, Switzerland and Sweden. In a statement published by UNI Global, the alliance said it would "demand fundamental human rights for all workers in Alphabet operations" and generally hold the company, which owns Google and various'other bets' including DeepMind, Waymo, Verily, Sidewalk Labs and Wing, accountable. "The problems at Alphabet--and created by Alphabet--are not limited to any one country, and must be addressed on a global level," said Christy Hoffman, UNI's General Secretary.

Fox News Flash top entertainment and celebrity headlines are here. Check out what's clicking today in entertainment. Disney is defending the host of a new "Star Wars" web series amid backlash to tweets some deemed to be racist toward White people. Krystina Arielle announced this month that she will host "The High Republic Show," a web series offering news and insights into the latest multimedia subseries of the immensely popular science fiction franchise. However, shortly after announcing Arielle as the host of the new bi-monthly show, some combed through her past tweets and found several that spoke in somewhat harsh terms about White people's role in dismantling racism.

The concepts of blameworthiness and wrongness are of fundamental importance in human moral life. But to what extent are humans disposed to blame artificially intelligent agents, and to what extent will they judge their actions to be morally wrong? To make progress on these questions, we adopted two novel strategies. First, we break down attributions of blame and wrongness into more basic judgments about the epistemic and conative state of the agent, and the consequences of the agent's actions. In this way, we are able to examine any differences between the way participants treat artificial agents in terms of differences in these more basic judgments. our second strategy is to compare attributions of blame and wrongness across human, artificial, and group agents (corporations). Others have compared attributions of blame and wrongness between human and artificial agents, but the addition of group agents is significant because these agents seem to provide a clear middle-ground between human agents (for whom the notions of blame and wrongness were created) and artificial agents (for whom the question remains open).

In a year that has seen decades' worth of global shocks, bad news, and scandals squeezed into 12 excruciatingly long months, the summer already feels like a distant memory. In August 2020, the world was in the throes of a major social and racial justice movement, and I argued hopefully in VentureBeat that the term "ethical AI" was finally starting to mean something. It was not the observation of a disinterested observer but an optimistic vision for coalescing the ethical AI community around notions of power, justice, and structural change. Yet in the intervening months it has proven to be, at best, an overly simplistic vision, and at worst, a naive one. The piece critiqued "second wave" ethical AI as being preoccupied with technical fixes to problems of bias and fairness in machine learning.

Google's testimony to an Australian Senate committee on Friday threatening to withdraw its search services from Australia is chilling to anyone who cares about democracy. It marks the latest escalation in the globally significant effort to regulate the way the big tech platforms use news content to drive their advertising businesses and the catastrophic impact on the news media across the world. The news bargaining code, which would require Google and Facebook to negotiate a fair price for the use of news content, is the product of an 18-month process driven by the competition regulator. That legislation is currently before the Australian parliament, where a Senate committee is taking final submissions from interested parties. The Google bombshell makes explicit what has been a slowly escalating threat that a binding code would not be tenable.

Accountability is widely understood as a goal for well governed computer systems, and is a sought-after value in many governance contexts. But how can it be achieved? Recent work on standards for governable artificial intelligence systems offers a related principle: traceability. Traceability requires establishing not only how a system worked but how it was created and for what purpose, in a way that explains why a system has particular dynamics or behaviors. It connects records of how the system was constructed and what the system did mechanically to the broader goals of governance, in a way that highlights human understanding of that mechanical operation and the decision processes underlying it. We examine the various ways in which the principle of traceability has been articulated in AI principles and other policy documents from around the world, distill from these a set of requirements on software systems driven by the principle, and systematize the technologies available to meet those requirements. From our map of requirements to supporting tools, techniques, and procedures, we identify gaps and needs separating what traceability requires from the toolbox available for practitioners. This map reframes existing discussions around accountability and transparency, using the principle of traceability to show how, when, and why transparency can be deployed to serve accountability goals and thereby improve the normative fidelity of systems and their development processes.