Three major players in the legal services market have created the first truly automated end-to-end digital solution for insurers, solicitors and counsel using Artificial Intelligence (AI). Keoghs Solicitors, St John's Buildings barristers' chambers and Advanced, the legal software specialist, have developed a digital solution using AI which enables Road Traffic Accident (RTA) personal injury cases to be litigated electronically, or identified as requiring a barrister, without the need for human intervention. The system has been in operation since late 2018 and is delivering a rapid, seamless and cost-saving service for clients. Keoghs created Lauri, the first AI litigation product, in 2017. Lauri reads unstructured data and pre-loads personal injury cases at Keoghs.

Although artificial intelligence (AI) has been around for decades, the industry has seen a recent resurgence. According to a report from PwC and CB Insights, AI investment in the US resulted in a 36 per cent compounded annual growth rate (CAGR) between 2013 and 2018. This jump in AI spending was more pronounced in 2018, increasing 72 per cent from the year prior. As momentum grows, intelligent software is permeating a diverse range of industries. The pursuit of operational efficiency and resulting cost savings remain primary drivers of AI adoption.

Bias in training data and proxy attributes are probably the main reasons for unfair machine learning outcomes. ML models are trained on historical data that are problematic due to the inherent societal bias. Besides, collecting labeled data in societal applications is challenging and costly. Subsequently, proxy attributes are often used as alternatives to labels. Yet, biased proxies cause model unfairness. In this paper, we introduce fair active learning (FAL) as a resolution. Considering a limited labeling budget, FAL carefully selects data points to be labeled in order to balance the model performance and fairness. Our comprehensive experiments on real datasets, confirm a significant fairness improvement while maintaining the model performance.

Last March, at the South by Southwest tech conference in Austin, Texas, Tesla and SpaceX founder Elon Musk issued a friendly warning: "Mark my words," he said, billionaire casual in a furry-collared bomber jacket and days' old scruff, "AI is far more dangerous than nukes." No shrinking violet, especially when it comes to opining about technology, the outspoken Musk has repeated a version of these artificial intelligence premonitions in other settings as well. "I am really quite close… to the cutting edge in AI, and it scares the hell out of me," he told his SXSW audience. "It's capable of vastly more than almost anyone knows, and the rate of improvement is exponential." Musk, though, is far from alone in his exceedingly skeptical (some might say bleakly alarmist) views. A year prior, the late physicist Stephen Hawking was similarly forthright when he told an audience in Portugal that AI's impact could be cataclysmic unless its rapid development is strictly and ethically controlled.

Whilst the main focus of Noble (2018) is on Google's algorithms and how its search engine reinforces racism and sexism, a wider social issue that can be deduced from her book, is the growth of artificial technology which might have negative consequences on the society, despite the assistance they provide in our everyday life. Noble (2018) believes that artificial intelligence will become a major human rights issue in the twenty-first century. Noble has spent some time debating the reasons for Google's racist and sexist search engine results. Her reasons range from paid advertising (I discussed this in my first blog post) to artificial errors developed by Google's automatic algorithms. These artificial errors can have serious repercussions on women because as pointed out by Halavais, search engines often help us with everyday life enquires and, therefore, we often trust in the results that appear without questioning it (As cited in Noble, 2018, p. 25).

Cerner was interviewing Silicon Valley giants to pick a storage provider for 250 million health records, one of the largest collections of U.S. patient data. Google dispatched former chief executive Eric Schmidt to personally pitch Cerner over several phone calls and offered around $250 million in discounts and incentives, people familiar with the matter say. Google had a bigger goal in pushing for the deal than dollars and cents: a way to expand its effort to collect, analyze and aggregate health data on millions of Americans. Google representatives were vague in answering questions about how Cerner's data would be used, making the health-care company's executives wary, the people say. Eventually, Cerner struck a storage deal with Amazon.com The failed Cerner deal reveals an emerging challenge to Google's move into health care: gaining the trust of health care partners and the public.

To support human decision making with machine learning models, we often need to elucidate patterns embedded in the models that are unsalient, unknown, or counterintuitive to humans. While existing approaches focus on explaining machine predictions with real-time assistance, we explore model-driven tutorials to help humans understand these patterns in a training phase. We consider both tutorials with guidelines from scientific papers, analogous to current practices of science communication, and automatically selected examples from training data with explanations. We use deceptive review detection as a testbed and conduct large-scale, randomized human-subject experiments to examine the effectiveness of such tutorials. We find that tutorials indeed improve human performance, with and without real-time assistance. In particular, although deep learning provides superior predictive performance than simple models, tutorials and explanations from simple models are more useful to humans. Our work suggests future directions for human-centered tutorials and explanations towards a synergy between humans and AI.

Algorithmic fairness in decision-making has been studied extensively in static settings where one-shot decisions are made on tasks such as classification. However, in practice most decision-making processes are of a sequential nature, where decisions made in the past may have an impact on future data. This is particularly the case when decisions affect the individuals or users generating the data used for future decisions. In this survey, we review existing literature on the fairness of data-driven sequential decision-making. We will focus on two types of sequential decisions: (1) past decisions have no impact on the underlying user population and thus no impact on future data; (2) past decisions have an impact on the underlying user population and therefore the future data, which can then impact future decisions. In each case the impact of various fairness interventions on the underlying population is examined.

We present a unifying framework encompassing many social choice settings. Viewing each social choice setting as voting in a suitable metric space, we consider a general model of social choice over metric spaces, in which---similarly to the spatial model of elections---each voter specifies an ideal element of the metric space. The ideal element functions as a vote, where each voter prefers elements that are closer to her ideal element. But it also functions as a proposal, thus making all participants equal not only as voters but also as proposers. We consider Condorcet aggregation and a continuum of solution concepts, ranging from minimizing the sum of distances to minimizing the maximum distance. We study applications of the abstract model to various social choice settings, including single-winner elections, committee elections, participatory budgeting, and participatory legislation. For each setting, we compare each solution concept to known voting rules and study various properties of the resulting voting rules. Our framework provides expressive aggregation for a broad range of social choice settings while remaining simple for voters, and may enable a unified and integrated implementation for all these settings, as well as unified extensions such as sybil-resiliency, proxy voting, and deliberative decision making.

Content analysis of AI patents shows the direction of the technology. If you want to know where a technology is going, follow the patents. That's what researchers from Stanford University and Brookings Institution recently did with artificial intelligence-related patents submitted to the US Patent and Trademark Office, and they have come up with an interesting read on the types of disruptions the technology will be having on jobs and companies. By analyzing keywords in AI patents, the researchers identified where AI will be having its greatest impact in the years ahead, above and beyond the "big--and often vague--claims" made about AI. "The technology remains a fluid and emergent topic, with no single definition and relatively little real-world examples of adoption to learn from," state Mark Muro, Jacob Whiton, and Robert Maxim, all with Brookings institution, who provide analysis on Stanford University Ph.D. candidate Michael Webb's approach of "quantifying the overlap between the text of AI patents and the text of job descriptions" that provides a unique way "to identify the kinds of tasks and occupations likely to be affected by particular AI capabilities." Such analysis, Muro and his team write, provides hard evidence on the direction AI will be taking in business settings.