English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In the long term, reinforcement learning (RL) is considered by many AItheorists to be the most promising path to artificial general intelligence.This places RL practitioners in a position to design systems that have neverexisted before and lack prior documentation in law and policy. Public agenciescould intervene on complex dynamics that were previously too opaque todeliberate about, and long-held policy ambitions would finally be madetractable. In this whitepaper we illustrate this potential and how it might betechnically enacted in the domains of energy infrastructure, social mediarecommender systems, and transportation. Alongside these unprecedentedinterventions come new forms of risk that exacerbate the harms alreadygenerated by standard machine learning tools. We correspondingly present a newtypology of risks arising from RL design choices, falling under fourcategories: scoping the horizon, defining rewards, pruning information, andtraining multiple agents. Rather than allowing RL systems to unilaterallyreshape human domains, policymakers need new mechanisms for the rule of reason,foreseeability, and interoperability that match the risks these systems pose.We argue that criteria for these choices may be drawn from emerging subfieldswithin antitrust, tort, and administrative law. It will then be possible forcourts, federal and state agencies, and non-governmental organizations to playmore active roles in RL specification and evaluation. Building on the "modelcards" and "datasheets" frameworks proposed by Mitchell et al. and Gebru etal., we argue the need for Reward Reports for AI systems. Reward Reports areliving documents for proposed RL deployments that demarcate design choices.

Choices, Risks, and Reward Reports: Charting Public Policy for Reinforcement Learning Systems