Artificial Intelligence: A Full-Spectrum Regulatory Challenge (Working Draft)

Artificial Intelligence: A Full-Spectrum Regulatory Challenge is a policy brief (working draft) authored by Sunil Abraham, with research contributions from Shweta Mohandas, Mira Swaminathan, and Shweta Reddy. Published by the Centre for Internet and Society (CIS) in October 2019, it presents a structured, domain-specific approach to regulating Artificial Intelligence (AI) in India and beyond.

The paper rejects one-size-fits-all ethical frameworks and omnibus AI laws. Instead, it argues for contextual regulation based on the actor (state, private sector, or individual), the degree of harm, and the human rights involved. It outlines how policymakers can deploy the entire spectrum of tools — from deregulation and forbearance to absolute prohibition — to balance innovation with accountability.

Editorial note: This page represents a working draft prepared during archival review and was not published as a final version.

Publication Details
Abstract
Editorial Status
Context and Background
Key Themes or Findings
Full Text
Collaborators and Acknowledgements
Citation

Publication Details

👤 Authors:: Sunil Abraham, Shweta Mohandas, Mira Swaminathan, and Shweta Reddy
🏛️ Published by:: The Centre for Internet and Society (CIS)
📅 Date:: 1 October 2019
📘 Type:: Policy Brief / Regulatory Framework for Artificial Intelligence
📄 Access:: Download PDF

Abstract

This policy brief articulates a full-spectrum approach to AI regulation, recognising that both over-regulation and under-regulation pose risks. It argues that governments must tailor regulation to the specific use, context, and harm potential of AI systems.

Key propositions include:

Using proportional regulatory tools across the spectrum — from self-regulation to prohibition.
Differentiating between state, corporate, and individual AI use.
Recognising that certain AI applications (e.g. mass surveillance) merit prohibition, while others (e.g. welfare delivery) require transparency and oversight.
Framing AI regulation as both a legal and ethical imperative.

Editorial Status

This page was prepared as part of an internal review and archival process and remained a working draft. It was not released as a final, published version on the site.

Working drafts are retained where useful to document the evolution of records, source verification, or editorial decisions. Content, structure, or supporting sections may therefore differ from the final published version or from later archival standards.

Context and Background

At the time of publication, global debates on AI ethics often prioritised voluntary principles over enforceable legal mechanisms. This brief, part of CIS’s policy series, bridges that gap by positioning AI governance as a multi-level regulatory challenge.

The framework builds on principles from data protection, accountability, competition, and human rights law. It also acknowledges India’s unique regulatory environment, suggesting pragmatic approaches that balance innovation with citizen protection.

Key Themes or Findings

Granular Regulation:
AI should be governed by a contextual matrix — actor, impact, and rights affected.
Law Enforcement and Surveillance:
Absolute prohibition on centralised mass surveillance; mandate Human-in-the-Loop (HITL) decision systems for all other uses.
Governance and Welfare:
Adopt Human-on-the-Loop (HOTL) for public service delivery; enforce FOSS, open standards, and open data mandates to ensure transparency.
Protected Domains:
Apply discrimination law, privacy impact assessments, and strict liability regimes for harm in sectors like healthcare, credit, and employment.
AI-Enabled Goods:
Require kill switches, safety standards, and right-to-tinker provisions; ensure insurance-backed liability for physical AI devices.
Social Media and Marketplaces:
Introduce user interface mandates and competition rules for algorithmic transparency and interoperability; encourage co-regulation.
Research and Development:
Bring AI research under Institutional Review Boards (IRBs); ensure ongoing ethical evaluation throughout project lifecycles.
Ethics and Individuals:
Ethics must complement, not replace, enforceable regulation.

Full Text

📄 Download Full Report (PDF)

Collaborators and Acknowledgements

Research Team: Shweta Mohandas, Mira Swaminathan, and Shweta Reddy
Interns: Vishnu Ramachandran, Chandramani Sirothia, Gursimar Setia, and Jwalika Balaji
Design: Saumyaa Naidu

Acknowledgements: Nalini Bharatula, Malavika Raghavan, Beni Chug, Mansi Kedia, Ujjwal Krishna, Isha Suri, Alok Prasanna Kumar, Divij Joshi, Arindrajit Basu, Amber Sinha, and Anubha Sinha.

Citation

APA

Abraham, S., Mohandas, S., Swaminathan, M., & Reddy, S. (2019).
Artificial Intelligence: A Full-Spectrum Regulatory Challenge.
The Centre for Internet and Society (CIS).
https://sunilabraham.in/publications/artificial-intelligence-full-spectrum/

BibTeX

@misc{abraham2019aispectrum,
  author = {Abraham, Sunil and Mohandas, Shweta and Swaminathan, Mira and Reddy, Shweta},
  title = {Artificial Intelligence: A Full-Spectrum Regulatory Challenge},
  year = {2019},
  howpublished = {The Centre for Internet and Society (CIS)},
  url = {https://sunilabraham.in/publications/artificial-intelligence-full-spectrum/}
}

MLA

Abraham, Sunil, Shweta Mohandas, Mira Swaminathan, and Shweta Reddy.
"Artificial Intelligence: A Full-Spectrum Regulatory Challenge."
The Centre for Internet and Society (CIS), 2019.
https://sunilabraham.in/publications/artificial-intelligence-full-spectrum/