The 3rd chapter of the textbook provides an overview of ethical issues that exist in the field of artificial intelligence. It provides cautionary tales, unintended consequences, and ethical considerations. It also covers biases that cause ethical issues and some tools that can help address them. We've spent many weeks writing the questionnaires. And the reason for that, is because we tried to think about what we wanted you to take away from each chapter.

Over the last few weeks, regulators and lawmakers around the world have made one thing clear: New laws will soon shape how companies use artificial intelligence (AI). In late March, the five largest federal financial regulators in the United States released a request for information on how banks use AI, signaling that new guidance is coming for the finance sector. Just a few weeks after that, the U.S. Federal Trade Commission (FTC) released an uncharacteristically bold set of guidelines on "truth, fairness, and equity" in AI -- defining unfairness, and therefore the illegal use of AI, broadly as any act that "causes more harm than good." The European Commission followed suit on April 21 released its own proposal for the regulation of AI, which includes fines of up to 6% of a company's annual revenues for noncompliance -- fines that are higher than the historic penalties of up to 4% of global turnover that can be levied under the General Data Protection Regulation (GDPR). For companies adopting AI, the dilemma is clear: On the one hand, evolving regulatory frameworks on AI will significantly impact their ability to use the technology; on the other, with new laws and proposals still evolving, it can seem like it's not yet clear what companies can and should do.

In 1949, at the dawn of the computer age, the French philosopher Gabriel Marcel warned of the danger of naively applying technology to solve life's problems. Life, Marcel wrote in Being and Having, cannot be fixed the way you fix a flat tire. Any fix, any technique, is itself a product of that same problematic world, and is therefore problematic, and compromised. Marcel's admonition is often summarized in a single memorable phrase: "Life is not a problem to be solved, but a mystery to be lived." Despite that warning, seventy years later, artificial intelligence is the most powerful expression yet of humans' urge to solve or improve upon human life with computers. But what are these computer systems? As Marcel would have urged, one must ask where they come from, whether they embody the very problems they would purport to solve. Ethics in AI is essentially questioning, constantly investigating, and never taking for granted the technologies that are being rapidly imposed upon human life. That questioning is made all the more urgent because of scale. AI systems are reaching tremendous size in terms of the compute power they require, and the data they consume. And their prevalence in society, both in the scale of their deployment and the level of responsibility they assume, dwarfs the presence of computing in the PC and Internet eras. At the same time, increasing scale means many aspects of the technology, especially in its deep learning form, escape the comprehension of even the most experienced practitioners. Ethical concerns range from the esoteric, such as who is the author of an AI-created work of art; to the very real and very disturbing matter of surveillance in the hands of military authorities who can use the tools with impunity to capture and kill their fellow citizens. Somewhere in the questioning is a sliver of hope that with the right guidance, AI can help solve some of the world's biggest problems. The same technology that may propel bias can reveal bias in hiring decisions. The same technology that is a power hog can potentially contribute answers to slow or even reverse global warming. The risks of AI at the present moment arguably outweigh the benefits, but the potential benefits are large and worth pursuing. As Margaret Mitchell, formerly co-lead of Ethical AI at Google, has elegantly encapsulated, the key question is, "what could AI do to bring about a better society?" Mitchell's question would be interesting on any given day, but it comes within a context that has added urgency to the discussion. Mitchell's words come from a letter she wrote and posted on Google Drive following the departure of her co-lead, Timnit Gebru, in December.

As practice group leader of a large (70 -lawyer) e-discovery group, I am frequently asked whether artificial intelligence will replace human lawyers. My favorite answer is "No – but lawyers who use AI will be replacing those who don't." Full disclosure – I did not come up with that answer. The first time I heard it was at a Cowen Café event, hosted by The Cowen Group, featuring thought leaders from in-house legal departments, outside law firms, and legal solution providers. While it is a pithy answer, I think it is probably accurate.

In what could be a harbinger of the future regulation of artificial intelligence (AI) in the United States, the European Commission published its recent proposal for regulation of AI systems. The proposal is part of the European Commission's larger European strategy for data, which seeks to "defend and promote European values and rights in how we design, make and deploy technology in the economy." To this end, the proposed regulation attempts to address the potential risks that AI systems pose to the health, safety, and fundamental rights of Europeans caused by AI systems. Under the proposed regulation, AI systems presenting the least risk would be subject to minimal disclosure requirements, while at the other end of the spectrum AI systems that exploit human vulnerabilities and government-administered biometric surveillance systems are prohibited outright except under certain circumstances. In the middle, "high-risk" AI systems would be subject to detailed compliance reviews.

The human voice conveys unique characteristics of an individual, making voice biometrics a key technology for verifying identities in various industries. Despite the impressive progress of speaker recognition systems in terms of accuracy, a number of ethical and legal concerns has been raised, specifically relating to the fairness of such systems. In this paper, we aim to explore the disparity in performance achieved by state-of-the-art deep speaker recognition systems, when different groups of individuals characterized by a common sensitive attribute (e.g., gender) are considered. In order to mitigate the unfairness we uncovered by means of an exploratory study, we investigate whether balancing the representation of the different groups of individuals in the training set can lead to a more equal treatment of these demographic groups. Experiments on two state-of-the-art neural architectures and a large-scale public dataset show that models trained with demographically-balanced training sets exhibit a fairer behavior on different groups, while still being accurate. Our study is expected to provide a solid basis for instilling beyond-accuracy objectives (e.g., fairness) in speaker recognition.

An essential part of argumentation-based reasoning is to identify arguments in favor and against a statement or query, select the acceptable ones, and then determine whether or not the original statement should be accepted. We present here an abstract framework that considers two independent forms of argument interaction--support and conflict--and is able to represent distinctive information associated with these arguments. This information can enable additional actions such as: (i) a more in-depth analysis of the relations between the arguments; (ii) a representation of the user's posture to help in focusing the argumentative process, optimizing the values of attributes associated with certain arguments; and (iii) an enhancement of the semantics taking advantage of the availability of richer information about argument acceptability. Thus, the classical semantic definitions are enhanced by analyzing a set of postulates they satisfy. Finally, a polynomial-time algorithm to perform the labeling process is introduced, in which the argument interactions are considered.

A series of recent developments points towards auditing as a promising mechanism to bridge the gap between principles and practice in AI ethics. Building on ongoing discussions concerning ethics-based auditing, we offer three contributions. First, we argue that ethics-based auditing can improve the quality of decision making, increase user satisfaction, unlock growth potential, enable law-making, and relieve human suffering. Second, we highlight current best practices to support the design and implementation of ethics-based auditing: To be feasible and effective, ethics-based auditing should take the form of a continuous and constructive process, approach ethical alignment from a system perspective, and be aligned with public policies and incentives for ethically desirable behaviour. Third, we identify and discuss the constraints associated with ethics-based auditing. Only by understanding and accounting for these constraints can ethics-based auditing facilitate ethical alignment of AI, while enabling society to reap the full economic and social benefits of automation.

Current large-scale language models can be politically biased as a result of the data they are trained on, potentially causing serious problems when they are deployed in real-world settings. In this paper, we describe metrics for measuring political bias in GPT-2 generation and propose a reinforcement learning (RL) framework for mitigating political biases in generated text. By using rewards from word embeddings or a classifier, our RL framework guides debiased generation without having access to the training data or requiring the model to be retrained. In empirical experiments on three attributes sensitive to political bias (gender, location, and topic), our methods reduced bias according to both our metrics and human evaluation, while maintaining readability and semantic coherence.

Ireland doesn't have a specific legal regime applicable to AI, yet it is widely deployed across a myriad of industries. We currently look to existing laws and try and fit AI into them. So when we want to incorporate AI applications into our businesses,we look for guidance in the strangest of places - the Employment Equality Acts 1988-2015 for HR deployments; the case of Donoghue -v- and Stephenson from 1932, when looking at general liability; and even the Control of Dogs Act 1986, when trying to find some analogies as to how robots or mobile AI might be treated! On 21 April the EU published its long awaited proposal for an AI Regulation. It now must go through the European Parliament and Council for consideration.