Control and Responsible Innovation in the Development of Autonomous Machines
Principal investigator: Wendell Wallach
Hastings investigator: Gregory E. Kaebnick
Funder: Future of Life Institute
As ever more sophisticated and autonomous computer networks and robots enter the commerce of daily life, scientists, entrepreneurs, and social theorists are confronted with a vast array of challenges to ensure that these systems are demonstrably beneficial, safe, controllable, and sensitive to human values. The developers of these technologies face questions about whether the machines can be designed and engineered to take the kind of morally significant actions that until now have been reserved for human beings. For example: What kinds of reasoning can and should they employ? Can autonomous systems be adequately controlled, and who will be responsible for their choices and actions? How can we meet challenges posed by potential future machines whose superintelligence will far exceed that of humans? Although many of the technologies are already upon us, we don’t yet have answers to these important questions.
The Hastings Center convened three solution-directed workshops with national and international experts that brought together leading figures in the many fields of research that will contribute to the safety, beneficially and controllability of AI and robotics. While other initiatives have approached the challenge of managing the development of AI in a piecemeal fashion, this project was the first to tackle the challenge comprehensively by encompassing safety procedures, creative and far-sighted engineering approaches, and legal and ethical oversight.
Recommendations from the project were presented at two public events: one in New York City and the other at Yale University. Watch the New York City event here.
Read the full recommendations here. Read the executive summary here.