Overview
As a result of recent advances in Large Language Models (LLMs), Artificial Intelligence (AI) has become a focus of popular discussion. Risks associated with AI have been considered for as long as such technologies have been imagined, and have been considered from a wide variety of perspectives. As such, there has been no shortage of discourse on the matter and there are now numerous calls to consider regulation, ethical frameworks, and even full halts to continued research on AI. Here we argue (i) that despite the very real risks associated with AI technologies, blanket regulation of and ethical frameworks for the broad range of AI technologies are inappropriate and likely to generate negative externalities and new conflicts, (ii) that instead, facilitation of amendment and adaptation of adjacent regulatory and self-regulatory systems, and instantiation of new professionalization, insurance, and self-regulatory structures would be far more productive, practical, and safe, and (iii) that the National Telecommunications and Information Administration (NTIA) is uniquely positioned to perform such facilitation and related convening and recommendation, given its mission and history. We conclude with summary recommendations.
Overview source: Curated paper metadata.
Methods and contributions
Read the source for the full argument, qualifications, and evidence.
Findings and contributions
- The authors argue blanket regulation and ethical frameworks for AI as a whole are inappropriate and likely to generate negative externalities and new conflicts.
- They argue sectoral standards (e.g. 'AI in medicine') built on existing licensure and duties of care are more practical than standards for AI as a whole.
- The comment describes HIPPO, a no-fault mutual self-insurance structure considered in forthcoming proposals by several contributors, with strict liability, pooled reserve and mass contract components.
- They call for data 'meta-standards' covering stable external reference to data and abstract entities, verified clearinghouses/data trusts, and annotation affordances.
- They recommend NTIA act as facilitating and convening authority for professionalization, data standardization/stable reference, and insurance-like structures.
Methods
- Joint public comment to NTIA docket NTIA-2023-0005-0001 — Four organizations (IRSIRI, AII, PFH, COGSEC) submitted a policy comment organized as background/basis followed by recommendations.
- BOLTS (business, operations, legal, technical, social) framing — The authors frame AI use-cases and risks across business, operations, legal, technical and social contexts drawing on their combined backgrounds.
Summary sources: Paper metadata and evidence · Extracted source text.
PDF downloads
Archived files available directly from this site.
- 2023_AIAccountability.pdf PDF · 0.53 MiB · extracted-text source
Citation
Citation metadata follows the unified bibliography.
Catalog details and resources
Related in Cognitive Security
Other catalogued works in the same domain.