Plenty of people are following the final days of the midterm election campaigns. Yale law researcher Rebecca Crootof has a special interest--a small wager. If she wins, victory will be bitter sweet, like the Manhattan cocktail that will be her prize. In June, Crootof bet that before 2018 is out an electoral campaign somewhere in the world will be roiled by a deepfake--a video generated by machine learning software that shows someone doing or saying something that in fact they did not do or say. Under the terms of the bet, the video must receive more than 2 million views before being debunked.

Public Knowledge, a Washington, DC advocacy group, today released a paper calling for the formation of a new federal government authority to develop expertise and capacity on artificial intelligence (AI), to be able to effectively regulate and govern these technologies in the future. The paper, The Inevitability of AI Law & Policy: Preparing Government for the Era of Autonomous Machines, was released on 31 October and written by Ryan Clough, general counsel at Public Knowledge. The paper "argues that the rapid and pervasive rise of artificial intelligence risks exploiting the most marginalized and vulnerable in our society." In order to mitigate these harms, a new national authority on AI will be needed, it says, to "provide the rest of the government with the expertise and experience needed to achieve five goals crucial to building ethical AI systems:

The International Conference on Advances in Computer Entertainment Technology, a relatively niche academic symposium in its 15th year, is embroiled in a white-hot controversy over its keynote speaker: Steve Bannon, the former White House Chief Strategist and founding member of the right-wing publication Breitbart. According to the conference organizer, Bannon--who has no academic background in computer science or interactive design but whose policy ideas have been embraced by white nationalists--will give a speech about how he believes "economic nationalism" will allow for a higher number of minorities to get jobs in sectors like computer science and gaming. The conference, also known as ACE, is scheduled to take place at the University of Montana in December. Since Bannon was added to the conference's roster last week, academics, scholarly associations, and university departments around the world have called for boycotting the conference, including the Milieux Institute for Arts, Culture and Technology at Concordia University, the Canadian Game Studies Association, and the Australian Digital Games Research Association. "Nothing of what Bannon can say represents the ACE community, or the games research community at large. His is a marginal discourse that should stay where it is, marginalized. And that's why we ask our community to #boycottACE," Miguel Angel Sicart, a games, art and interactive design researcher at the IT University of Copenhagen, told WIRED in an email.

China has released a new propaganda video of its deadly unmanned fighter jet, which shows the aircraft striking still and moving targets. CH-5, also known as Rainbow-5, was unveiled in 2016 and is said to be China's largest and most powerful drone bomber. The aircraft can carry 16 missiles and strike targets while flying at an altitude of 6,000 metres (19,685 feet), Chinese media have claimed. It can fly up to 60 hours without refuelling with a maximum flight altitude of 8,000 metres (26,246 feet) and a maximum range of 10,000 kilometres (6,213 miles). CH-5, also known as Rainbow-5, is on display during the 11th China International Aviation and Aerospace Exhibition in 2016.

Unsupervised node embedding methods (e.g., DeepWalk, LINE, and node2vec) have attracted growing interests given their simplicity and effectiveness. However, although these methods have been proved effective in a variety of applications, none of the existing work has analyzed the robustness of them. This could be very risky if these methods are attacked by an adversarial party. In this paper, we take the task of link prediction as an example, which is one of the most fundamental problems for graph analysis, and introduce a data positioning attack to node embedding methods. We give a complete characterization of attacker's utilities and present efficient solutions to adversarial attacks for two popular node embedding methods: DeepWalk and LINE. We evaluate our proposed attack model on multiple real-world graphs. Experimental results show that our proposed model can significantly affect the results of link prediction by slightly changing the graph structures (e.g., adding or removing a few edges). We also show that our proposed model is very general and can be transferable across different embedding methods. Finally, we conduct a case study on a coauthor network to better understand our attack method.

Topic models are widely used for thematic structure discovery in text. But traditional topic models often require dedicated inference procedures for specific tasks at hand. Also, they are not designed to generate word-level semantic representations. To address these limitations, we propose a topic modeling approach based on Generative Adversarial Nets (GANs), called Adversarial-neural Topic Model (ATM). The proposed ATM models topics with Dirichlet prior and employs a generator network to capture the semantic patterns among latent topics. Meanwhile, the generator could also produce word-level semantic representations. To illustrate the feasibility of porting ATM to tasks other than topic modeling, we apply ATM for open domain event extraction. Our experimental results on the two public corpora show that ATM generates more coherence topics, outperforming a number of competitive baselines. Moreover, ATM is able to extract meaningful events from news articles.

Recently, many AI researchers and practitioners have embarked on research visions that involve doing AI for "Good". This is part of a general drive towards infusing AI research and practice with ethical thinking. One frequent theme in current ethical guidelines is the requirement that AI be good for all, or: contribute to the Common Good. But what is the Common Good, and is it enough to want to be good? Via four lead questions, I will illustrate challenges and pitfalls when determining, from an AI point of view, what the Common Good is and how it can be enhanced by AI. The questions are: What is the problem / What is a problem?, Who defines the problem?, What is the role of knowledge?, and What are important side effects and dynamics? The illustration will use an example from the domain of "AI for Social Good", more specifically "Data Science for Social Good". Even if the importance of these questions may be known at an abstract level, they do not get asked sufficiently in practice, as shown by an exploratory study of 99 contributions to recent conferences in the field. Turning these challenges and pitfalls into a positive recommendation, as a conclusion I will draw on another characteristic of computer-science thinking and practice to make these impediments visible and attenuate them: "attacks" as a method for improving design. This results in the proposal of ethics pen-testing as a method for helping AI designs to better contribute to the Common Good.

The term "artificial intelligence" gets thrown around a lot today, especially in government circles, where leaders are eager to get ahead of the technological curve. But the military, at least, doesn't really want true AI, according to a former deputy defense secretary. "We are not talking about Skynets and we're not talking about Terminators. Those are what you would call an artificial general intelligence-type weapon," Bob Work, who served as deputy defense secretary under Presidents Obama and Trump, said Tuesday during a speech at the annual SAP NS2 Solutions Summit. "We're looking for narrow AI systems that can compose courses of action to accomplish the tasks that the machine is given and it can choose among the courses of action."

VIENNA – Edmund Rath, a 53-year-old from Austria, wants to do simple things such as brush his teeth and slice bread. Most people take such activities for granted but not Rath, who lost his arm just below the shoulder in a truck accident last year that ended his career as a builder. That bad luck was followed by a stroke of fortune. He was chosen by Austrian surgeons as the first person to have a single operation in May to install a click-on prosthesis that the brain controls with signals to the missing hand. The procedure, known as osseointegration (OI), involved implanting a metal rod into the bone of his residual limb.

UK businesses are at risk of falling behind international rivals because of a failure to use more artificial intelligence, a report by Microsoft has warned. The computing giant found that more than half of UK leaders have no AI strategy in place within their organisation. More than a third admit they are not focused on AI at all, while studies suggest that organisations making use of AI are outperforming those who aren't by five per cent. UK businesses are at risk of falling behind international rivals because of a failure to use more artificial intelligence, a report by Microsoft has warned. Of the 1,000 business leaders surveyed in partnership with Goldsmiths, University of London and YouGov, four in ten said that they believe their current business model will cease to exist within five years.