Technology Law: Artificial Intelligence: Trust and Distrust
Robin Feldman · bepress Legal Repository · 2019
Artificial intelligence (AI) is percolating through modern society.In the automobile industry, AI systems assist drivers with steering, changing lanes, and parking.Early AI projects in the criminal-justice system predict where crime is likely to occur for the purpose of targeting policing.Smart glasses tailored to business applications are emerging into the marketplace.Eventually, these glasses will use machine learning to identify objects and voices, prompting the wearer to take certain actions or setting out a range of possible actions.Banking and insurance firms use AI to advise customers on financial services, assess consumer risk, and monitor for fraud.Employers use AI systems in hiring decisions.And in the healthcare field, invasive brain interfaces have demonstrated the ability for thought control of complex robotic limbs and virtual agents.As AI becomes a ubiquitous part of our everyday life, a key aspect will be the way in which society-and by extension, the legal systemmanages both the integration of these systems and society's expectations.Society will have to learn to trust the capacity of AI systems sufficiently so that it can soar to new heights, without succumbing to the "irrational exuberance" 2 that can send society crashing to the ground when AI fails to live up to unreasonable expectations.And society must learn to tolerate the ambiguity that lies between these two extremes. The State of AIAI refers to any artificial agent that ingests data and responds to that data by completing a task.The AI we have today is "narrow AI": AI that can complete only one or a handful of pre-specified tasks.We are nowhere near "Artificial General Intelligence-defined by its potentially unattainable ability to complete any task a human can, short