Content area
Full text
As artificial intelligence (AI) continues to embed itself in our daily lives, many focus on the threats it poses to privacy, security, due process, and democracy itself. But beyond these legitimate concerns, AI promises to optimize activities, increase efficiency, and enhance the accuracy and efficacy of the many aspects of society relying on predictions and likelihoods. In short, its most promising applications may come, not from uses affecting civil liberties and the social fabric of our society, but from those particularly complex technical problems lying beyond our ready human capacity. Climate change is one such complex problem, requiring fundamental changes to our transportation, agricultural, building, and energy sectors. This Article argues for the enhanced use of AI to address climate change, using the energy sector to exemplify its potential promise and pitfalls. The Article then analyzes critical policy tradeoffs that may be associated with an increased use of AI and argues for its disciplined use in a way that minimizes its limitations while harnessing its benefits to reduce greenhouse-gas emissions.
Introduction
Headlines claim that robots will take our jobs1 and that drones may use machine learning to teach themselves to launch an independent attack.2 Scholars have documented more immediate concerns about the bias surrounding the use of algorithms for criminal sentencing3 and facial recognition programs.4 Together, each of these scenarios implicates various forms of artificial intelligence (AI). An amorphous term that has moved from obscurity to commonplace as of late, AI here refers to "a set of techniques aimed at approximating some aspect of human or animal cognition using machines."5 There are many different forms of AI. The most popular form is machine learning, a technology used to make predictions that functions best when using massive amounts of data and computing capacity.6 AI's tentacles have pervaded many aspects of society, from more mundane uses in Google searches to more sophisticated uses in criminal-bail sentencing,7 autonomous vehicles,8 e-commerce, digital advertising, and medicine.9 Like most technological innovations, these techniques have the capacity for both beneficial and detrimental outcomes.10
This Article, prepared in connection with the Yale Journal on Regulation's symposium on Regulating the Technological Frontier, attempts to strike a more optimistic tone for AL It strives to remind its readers that some of the problems plaguing...