Positive, reliable and human-centric artificial intelligence (AI) relies on the willingness of Europe as a whole to design a balanced and inclusive governance framework that would allow it to become a leader in the development of trustworthy AI technologies worldwide. That was the main conclusion reached in the frame of the high-level workshop organised by the Panel for the Future of Science and Technology (STOA) on 29 January 2020 at the European Parliament in Brussels. The first STOA event for this parliamentary term (2019-2024) drew a full house with Members of the European Parliament, European Commission leaders, academic experts and representatives of international organisations debating how to strike the right balance on AI. Harnessing the numerous benefits that the transformative power of AI can bring needs to also take account of the necessity to mitigate a number of potential risks – from hampering people's fundamental rights, such as privacy or non-discrimination – to undermining European values such as democracy, human dignity and the right to assemble. The event proved to be a timely occasion to discuss how Europe could maximise the benefits and address the challenges of AI in a human-centric way, coming only a few days before the publication of the European Commission's legislative plans on AI in the form of a White Paper on 19 February 2020.

Today's global sanctions regimes have arguably never been more challenging for organisations to ensure they remain compliant and have the required screening processes and procedures in place. Over the past decade, trade and economic sanctions have become an ever more popular tool of foreign policy in an increasingly uncertain geo-political climate. Aside from country-specific sanctions, such as those against Iran, Russia, North Korea, etc, more targeted regulations focus upon particular businesses or individuals. As a result, national and international AML, screening and anti-fraud obligations have increased in both scope and complexity. Failure to comply with sanctions and money laundering obligations, can result in severe financial and reputational costs.

The future of artificial intelligence (AI) is here: self-driving cars, grocery-delivering drones and voice assistants like Alexa that control more and more of our lives, from the locks on our front doors to the temperatures of our homes. For example, should an autonomous vehicle swerve into a pedestrian or stay its course when facing a collision? These questions plague technology companies as they develop AI at a clip outpacing government regulation, and have led Seattle University to develop a new ethics course for the public. Launched last week, the free, online course for businesses is the first step in a Microsoft-funded initiative to merge ethics and technology education at the Jesuit university. Seattle U senior business-school instructor Nathan Colaner hopes the new course will become a well-known resource for businesses "as they realize that [AI] is changing things," he said.

Last decade has seen major improvements in the performance of artificial intelligence which has driven wide-spread applications. Unforeseen effects of such mass-adoption has put the notion of AI safety into the public eye. AI safety is a relatively new field of research focused on techniques for building AI beneficial for humans. While there exist survey papers for the field of AI safety, there is a lack of a quantitative look at the research being conducted. The quantitative aspect gives a data-driven insight about the emerging trends, knowledge gaps and potential areas for future research. In this paper, bibliometric analysis of the literature finds significant increase in research activity since 2015. Also, the field is so new that most of the technical issues are open, including: explainability with its long-term utility, and value alignment which we have identified as the most important long-term research topic. Equally, there is a severe lack of research into concrete policies regarding AI. As we expect AI to be the one of the main driving forces of changes in society, AI safety is the field under which we need to decide the direction of humanity's future.

The increasing deployment of machine learning as well as legal regulations such as EU's GDPR cause a need for user-friendly explanations of decisions proposed by machine learning models. Counterfactual explanations are considered as one of the most popular techniques to explain a specific decision of a model. While the computation of "arbitrary" counterfactual explanations is well studied, it is still an open research problem how to efficiently compute plausible and feasible counterfactual explanations. We build upon recent work and propose and study a formal definition of plausible counterfactual explanations. In particular, we investigate how to use density estimators for enforcing plausibility and feasibility of counterfactual explanations. For the purpose of efficient computations, we propose convex density constraints that ensure that the resulting counterfactual is located in a region of the data space of high density.

On January 1, 2020, Illinois' new Artificial Intelligence Video Interview Act (AIVIA) went into effect, meaning Illinois employers must now comply with the law if they use artificial intelligence (AI) to analyze video interviews by job candidates. As we outlined in a prior post, the AIVIA imposes duties of transparency, consent and data destruction on organizations using AI to evaluate interviewees for jobs that are "based" in Illinois. While these concepts may be clear in the abstract, the Illinois law is a lesson in brevity and leaves several key terms undefined (including, for example, the term "artificial intelligence"). Nor is it clear what it means for a position to be "based" in Illinois. As a result, employers using AI-enabled analytics in interview videos must sort through these questions and take other affirmative steps to ensure compliance with the new law.

What is artificial intelligence and why can it be dangerous? As learning algorithms can process data sets with precision and speed beyond human capacity, artificial intelligence (AI) applications have become increasingly common in finance, healthcare, education, the legal system and beyond. However, reliance on AI also carries risks, especially where decisions are made without human oversight. Machine learning relies on pattern-recognition within datasets. Problems arise when the available data reflects societal bias.

As legal pressures and US lawmaker scrutiny mounts, Clearview AI, the facial recognition company that claims to have a database of more than 3 billion photos scraped from websites and social media, is looking to grow around the world. A document obtained via a public records request reveals that Clearview has been touting a "rapid international expansion" to prospective clients using a map that highlights how it either has expanded, or plans to expand, to at least 22 more countries, some of which have committed human rights abuses. The document, part of a presentation given to the North Miami Police Department in November 2019, includes the United Arab Emirates, a country historically hostile to political dissidents, and Qatar and Singapore, the penal codes of which criminalize homosexuality. Clearview CEO Hoan Ton-That declined to explain whether Clearview is currently working in these countries or hopes to work in them. He did confirm that the company, which had previously claimed that it was working with 600 law enforcement agencies, has relationships with two countries on the map.

Predictive storage analytics, in conjunction with AI, is remaking enterprise data storage management. As more vendors incorporate these features into their products, storage is becoming a dynamic part of the IT infrastructure. This is a new world where technology does more than deliver an accurate analysis of problems to experts who must then figure out the next step. Instead, when combined with AI and machine learning, predictive storage analytics resolves issues with the storage infrastructure and takes proactive steps to prevent future problems. This reduces the need for IT staff intervention, eliminates downtime and saves time and money.

Data-driven predictive models are increasingly used to inform decisions that hold important consequences for individuals and society. As a result, decision makers are often obliged, even legally required, to provide explanations about their decisions. In this context, it has been increasingly argued that these explanations should help individuals understand what would have to change for these decisions to be beneficial ones. However, there has been little discussion on the possibility that individuals may use the above counterfactual explanations to invest effort strategically in order to maximize their chances of receiving a beneficial decision. In this paper, our goal is to find policies and counterfactual explanations that are optimal in terms of utility in such a strategic setting. To this end, we first show that, given a pre-defined policy, the problem of finding the optimal set of counterfactual explanations is NP-hard. However, we further show that the corresponding objective is nondecreasing and satisfies submodularity. Therefore, a standard greedy algorithm offers an approximation factor of $(1-1/e)$ at solving the problem. Additionally, we also show that the problem of jointly finding both the optimal policy and set of counterfactual explanations reduces to maximizing a non-monotone submodular function. As a result, we can use a recent randomized algorithm to solve the problem, which offers an approximation factor of $1/e$. Finally, we illustrate our theoretical findings by performing experiments on synthetic and real lending data.