This chapter argues that big data analytics, including machine learning and artificial intelligence, are natural outgrowths of recent developments in computer technology such as the availability of massive data sets, vast increases in computing power, and breakthroughs in analytical techniques. These techniques promise unprecedented benefits for consumers, workers, and society at large, but they also pose challenges for privacy and fairness. The chapter contains a short summary of the range of potential benefits made possible by these new analytic techniques and then discusses privacy and fairness challenges. Principles of privacy policy requiring data minimization and restricting secondary data use need to be reformulated to allow for both the successful delivery of big data benefits and effective privacy protection. Ubiquitous re-identification risks and information externalities reduce the ability of individuals to control the disclosure of information and suggest less reliance on notice and choice mechanisms. Big data analytics can pose fairness challenges, but these techniques are not exempt from existing antidiscrimination and consumer protection laws. Regulatory agencies and courts need to enforce these laws against any abuses accomplished through big data analysis. Disclosure of source code is not an effective way to respond to the challenges of designing and using unbiased algorithms. Instead, enterprises should develop and implement a framework for responsible use of data analytics that will provide for fairness by design and after-the-fact audits of algorithms in use. Such a framework will need to adopt standards of fairness and appropriate remedies for findings of disparate impact. This will require moving beyond technical matters to address sensitive normative issues where the interests of different groups collide and moral intuitions diverge. A collaborative effort of businesses, governments, academics, and civil rights and public interest groups might sharpen the issues and allow sharing of information and best practices in a way that would benefit all
