Inc. in June pitched its facial recognition technology -- which can identify people from surveillance footage using image databases -- as a tool for U.S. Immigration and Customs Enforcement, showing that Amazon continued to push the software to law enforcement agencies as criticism swirled from the company's workforce and civil liberties groups. Employees in the Amazon Web Services cloud-computing unit met with the federal agency in California to present its artificial intelligence tools, according to emails obtained by the nonprofit Project on Government Oversight. Those tools include Rekognition, which uses artificial intelligence to quickly identify people in photos and videos. The software enables law enforcement to track individuals from cameras in public places. The American Civil Liberties Union in May criticized the use of the technology by police departments in Oregon and Florida, saying it threatened civil rights.

As the No Free Lunch theorem formally states [1], algorithms for detecting communities in networks must make tradeoffs. In this work, we present a method for using metadata to inform tradeoff decisions. We extend the content map equation, which adds metadata entropy to the traditional map equation, by introducing a tuning parameter allowing for explicit specification of the metadata's relative importance in assigning community labels. On synthetic networks, we show how tuning for node metadata relates to the detectability limit, and on empirical networks, we show how increased tuning for node metadata yields increased mutual information with the metadata at a cost in the traditional map equation. Our tuning parameter, like the focusing knob of a microscope, allows users to "zoom in" and "zoom out" on communities with varying levels of focus on the metadata.

Machine learning is an important tool for decision making, but its ethical and responsible application requires rigorous vetting of its interpretability and utility: an understudied problem, particularly for natural language processing models. We design a task-specific evaluation for a question answering task and evaluate how well a model interpretation improves human performance in a human-machine cooperative setting. We evaluate interpretation methods in a grounded, realistic setting: playing a trivia game as a team. We also provide design guidance for natural language processing human-in-the-loop settings.

While the future of humankind is artificial intelligence, what exactly is the future of artificial intelligence? This is the insoluble ethical dilemma of artificial intelligence. These systems are autonomous and self-learning. There is no question that artificial intelligence will lead to many positive outcomes. However, the flip side is the question of ethics and the challenge of ensuring that machines driven by artificial intelligence will "behave" ethically towards living things.

As the ways to utilize artificial intelligence (AI) grow, so do the amount of people implementing this cutting-edge technology to their work. According to Statista, the rate of adoption is quite high -- global revenues from AI for enterprise applications is projected to grow from $1.62B in 2018 to $31.2B in 2025. Its popularity reaches every sector and is substantially affecting the legal space as well. The possibilities for lawyers to benefit from AI are endless. Many automated and machine-learning features enable lawyers to complete a number of tasks faster and cheaper.

A "tech tax" is necessary if the world is to avoid a dystopian future in which AI leads to a concentration of global wealth in the hands of a few thousand people, influential economist Dr Jeffrey Sachs has warned. Speaking to the Guardian, Sachs backed calls for taxation aimed at the largest tech companies, arguing that new technologies were dramatically shifting the income distribution worldwide "from labour to intellectual property (IP) and other capital income." "So rather than cutting capital income taxation, as we've been doing in a race to the bottom, we ought to be finding ways to tax capital income and IP income," Sachs added. "Things like the proposed tech tax are actually a very good idea. The specific form of it is debatable, but the idea is that five companies are worth $3.5tn, basically because of network externalities and information monopolies, and therefore are absolutely right for efficient taxation."

Between 2014 and 2017 Amazon tried to build an algorithmic system to analyze resumes and suggest the best hires. An anonymous Amazon employee called it the "holy grail" if it actually worked. After the company trained the algorithm on 10 years of its own hiring data, the algorithm reportedly became biased against female applicants. The word "women," like in women's sports, would cause the algorithm to specifically rank applicants lower. After Amazon engineers attempted to fix that problem, the algorithm still wasn't up to snuff and the project was ended.

Authors: Yan Li, University of Michigan Chandan K. Reddy, Department of Computer Science, Virginia Polytechnic Institute and State University Abstract: Due to the advancements in various data acquisition and storage technologies, different disciplines have attained the ability to not only accumulate a wide variety of data but also to monitor observations over longer time periods. In many real-world applications, the primary objective of monitoring these observations is to estimate when a particular event of interest will occur in the future. One of the major difficulties in handling such problem is the presence of censoring, i.e., the event of interests is unobservable in some instance which is either because of time limitation or losing track. Due to censoring, standard statistical and machine learning based predictive models cannot readily be applied to analyze the data. An important subfield of statistics called survival analysis provides different mechanisms to handle such censored data problems.

Technology companies should find effective channels of communication with local civil society groups and researchers, particularly in geographic areas where human rights concerns are high, in order to identify and respond to risks related to AI deployments. Technology companies and researchers should conduct Human Rights Impact Assessments (HRIAs) through the life cycle of their AI systems. Toolkits should be developed to assess specific industry needs. Governments should acknowledge their human rights obligations and incorporate a duty to protect fundamental rights in national AI policies, guidelines, and possible regulations. Governments can play a more active role in multilateral institutions, like the UN, to advocate for AI development that respects human rights.

Spatiotemporal sequence prediction is an important problem in deep learning. We study next-frame(s) video prediction using a deep-learning-based predictive coding framework that uses convolutional, long short-term memory (convLSTM) modules. We introduce a novel reduced-gate convolutional LSTM (rgcLSTM) architecture that requires a significantly lower parameter budget than a comparable convLSTM. Our reduced-gate model achieves equal or better next-frame(s) prediction accuracy than the original convolutional LSTM while using a smaller parameter budget, thereby reducing training time. We tested our reduced gate modules within a predictive coding architecture on the moving MNIST and KITTI datasets. We found that our reduced-gate model has a significant reduction of approximately 40 percent of the total number of training parameters and a 25 percent redution in elapsed training time in comparison with the standard convolutional LSTM model. This makes our model more attractive for hardware implementation especially on small devices.