Germany is the first country in the world to legalize fully autonomous vehicles on public roads. The landmark legislation passed both the lower and upper chambers of Germany's parliament last week with a comfortable majority. The legislation will allow Level 4 autonomous vehicles to operate on public roads in Germany without drivers behind the wheel and without obtaining special permits. The new legal framework mandates that autonomous vehicles must be manufactured and maintained in accordance with new, yet-to-be-crafted technical requirements. In addition, the law calls for technical command centers where live supervisors will oversee each fleet of self-driving cars and have the ability to control and deactivate them remotely when problems arise.

Securing networked infrastructures is important in the real world. The problem of deploying security resources to protect against an attacker in networked domains can be modeled as Network Security Games (NSGs). Unfortunately, existing approaches, including the deep learning-based approaches, are inefficient to solve large-scale extensive-form NSGs. In this paper, we propose a novel learning paradigm, NSG-NFSP, to solve large-scale extensive-form NSGs based on Neural Fictitious Self-Play (NFSP). Our main contributions include: i) reforming the best response (BR) policy network in NFSP to be a mapping from action-state pair to action-value, to make the calculation of BR possible in NSGs; ii) converting the average policy network of an NFSP agent into a metric-based classifier, helping the agent to assign distributions only on legal actions rather than all actions; iii) enabling NFSP with high-level actions, which can benefit training efficiency and stability in NSGs; and iv) leveraging information contained in graphs of NSGs by learning efficient graph node embeddings. Our algorithm significantly outperforms state-of-the-art algorithms in both scalability and solution quality.

The first generation of AI has picked up on human biases. Among many disturbing cases of biased AI systems resulting in discriminatory outcomes, the most heart-breaking ones were cases involving unfair elongation of prison sentence, unfair credit card decision, and home appraisal outcomes. So, how does bias get into AI systems? While this is by no means an excuse, it does point to the key problem -- almost no focus was given to ensuring the moral, social, and responsible aspect of AI- often termed Ethical AI. A 2019 Gartner study reported that by 2022, 30% of the companies will invest in explainable ethical AI, from almost none in 2019.

Artificial intelligence has had a huge impact on many industries in recent years and will continue to benefit them in the future. The pandemic-induced acceleration of technology adoption has led many sectors, both private and public to leverage AI for their advantage and growth. In the last few years, AI has enabled many innovations and driven the proliferation of technologies like IoT, robotics, analytics, and voice assistants. According to a report, AI topped the patent filings in 2020. This is not new, AI has been securing a large number of patents in the last few years.

Artificial Intelligence (AI) today is rapidly changing the face of technology. But with the ability to create devices and systems capable of autonomous decisions, arises the need for legislation to monitor AI. Amazon's now scrapped AI recruiting tool is a prime example where it was discovered that the AI tool had a bias towards men since it had been trained on 10 years of data when men held most tech positions. As we constantly move towards a more technology integrated world, the need for the right balance in legislation grows more important. It needs to protect the rights of the citizens alongside ensuring that it is not a hindrance to technology and business growth.

The federal government should establish an AI Safety Commissioner and halt the use of facial recognition and algorithms in important decision-making until adequate protections are in place, the Australian Human Rights Commission has concluded after a three-year investigation. The Australian Human Rights Commission's (AHRC) report on Human Rights and Technology was tabled in Parliament on Thursday afternoon, with 38 recommendations to the government on ensuring human rights are upheld in the laws, policies, funding and education on artificial intelligence. Human Rights Commissioner Ed Santow has urged local, state, territory and federal governments to put on hold the use of facial recognition and AI in decision-making that has a significant impact on individuals. This moratorium should be until adequate legislation is in place that regulates the use of these technologies and ensures human rights are protected. The use of automation and algorithms in government decision-making should also be paused until a range of protections and transparency measures are in place, Mr Santow said in the report.

Explaining the decisions of an Artificial Intelligence (AI) model is increasingly critical in many real-world, high-stake applications. Hundreds of papers have either proposed new feature attribution methods, discussed or harnessed these tools in their work. However, despite humans being the target end-users, most attribution methods were only evaluated on proxy automatic-evaluation metrics [52, 66, 68]. In this paper, we conduct the first, large-scale user study on 320 lay and 11 expert users to shed light on the effectiveness of state-of-the-art attribution methods in assisting humans in ImageNet classification, Stanford Dogs fine-grained classification, and these two tasks but when the input image contains adversarial perturbations. We found that, in overall, feature attribution is surprisingly not more effective than showing humans nearest training-set examples. On a hard task of fine-grained dog categorization, presenting attribution maps to humans does not help, but instead hurts the performance of human-AI teams compared to AI alone. Importantly, we found automatic attribution-map evaluation measures to correlate poorly with the actual human-AI team performance. Our findings encourage the community to rigorously test their methods on the downstream human-in-the-loop applications and to rethink the existing evaluation metrics.

The widespread utilization of AI systems has drawn attention to the potential impacts of such systems on society. Of particular concern are the consequences that prediction errors may have on real-world scenarios, and the trust humanity places in AI systems. It is necessary to understand how we can evaluate trustworthiness in AI and how individuals and entities alike can develop trustworthy AI systems. In this paper, we analyze each element of trustworthiness and provide a set of 20 guidelines that can be leveraged to ensure optimal AI functionality while taking into account the greater ethical, technical, and practical impacts to humanity. Moreover, the guidelines help ensure that trustworthiness is provable and can be demonstrated, they are implementation agnostic, and they can be applied to any AI system in any sector.

An automated system that could assist a judge in predicting the outcome of a case would help expedite the judicial process. For such a system to be practically useful, predictions by the system should be explainable. To promote research in developing such a system, we introduce ILDC (Indian Legal Documents Corpus). ILDC is a large corpus of 35k Indian Supreme Court cases annotated with original court decisions. A portion of the corpus (a separate test set) is annotated with gold standard explanations by legal experts. Based on ILDC, we propose the task of Court Judgment Prediction and Explanation (CJPE). The task requires an automated system to predict an explainable outcome of a case. We experiment with a battery of baseline models for case predictions and propose a hierarchical occlusion based model for explainability. Our best prediction model has an accuracy of 78% versus 94% for human legal experts, pointing towards the complexity of the prediction task. The analysis of explanations by the proposed algorithm reveals a significant difference in the point of view of the algorithm and legal experts for explaining the judgments, pointing towards scope for future research.

Group-fairness in classification aims for equality of a predictive utility across different sensitive sub-populations, e.g., race or gender. Equality or near-equality constraints in group-fairness often worsen not only the aggregate utility but also the utility for the least advantaged sub-population. In this paper, we apply the principles of Pareto-efficiency and least-difference to the utility being accuracy, as an illustrative example, and arrive at the Rawls classifier that minimizes the error rate on the worst-off sensitive sub-population. Our mathematical characterization shows that the Rawls classifier uniformly applies a threshold to an ideal score of features, in the spirit of fair equality of opportunity. In practice, such a score or a feature representation is often computed by a black-box model that has been useful but unfair. Our second contribution is practical Rawlsian fair adaptation of any given black-box deep learning model, without changing the score or feature representation it computes. Given any score function or feature representation and only its second-order statistics on the sensitive sub-populations, we seek a threshold classifier on the given score or a linear threshold classifier on the given feature representation that achieves the Rawls error rate restricted to this hypothesis class. Our technical contribution is to formulate the above problems using ambiguous chance constraints, and to provide efficient algorithms for Rawlsian fair adaptation, along with provable upper bounds on the Rawls error rate. Our empirical results show significant improvement over state-of-the-art group-fair algorithms, even without retraining for fairness.