Best Paper Award · 2026
ComSIA 2026
Springer, Lecture Notes in Networks and Systems. Awarded by the conference technical committee for research on governance of connected AI systems.
AI Governance Risk Architect + Speaker
Vivek
KumarGovern what AI can reach.
Vivek helps leaders move from AI ambition to controlled, auditable, enterprise-ready adoption.
Independent AI governance risk architect · Creator of PAARA
15+ yearsPrivacy, data protection, security and AI governance
Creator of PAARAConnector-layer governance for enterprise AI
AI2030 Global FellowMerit-selected global AI leadership fellowship
AIGP · FIP · CIPP/US · CIPP/E · CIPM · CISMCertified across AI governance, privacy and security
01 / Point of view
The governance gap is often in the connections.
A model can look controlled in isolation while its connectors quietly widen access, combine sensitive context, and enable action across the enterprise.
Vivek translates this technical reality into decisions leaders can own. His work focuses on the connector layer behind generative and agentic AI: permissions, retrieval boundaries, data paths, delegated actions, evidence, and accountable human oversight.
The aim is practical: turn governance from a policy artifact into an operating discipline that delivery teams can use and executives can defend.
See the PAARA method01
What can the system reach?
02
What can it combine?
03
What can it reveal?
04
What can it act on?
05
Who remains accountable?
02 / Signature talks
Three conversations for the agentic enterprise.
Keynotes and executive sessions built to sharpen decisions, clarify accountability, and give teams a practical way forward.
Talk 01
Govern What AI Can Reach
The Connector Layer Behind Enterprise AI Risk
Govern What AI Can Reach
The Connector Layer Behind Enterprise AI Risk
Enterprise AI assistants do not stay inside the model. They connect to email, documents, customer systems, knowledge bases, and tools, then assemble context across boundaries that traditional inventories rarely capture. Vivek gives leaders a five-question method for finding these hidden data-access and accountability risks before a deployment scales. The session replaces abstract warnings with a concrete operating view of permissions, retrieval, aggregation, logging, and control ownership.
Talk 02
From AI Policy to Operating Control
How Leaders Make Governance Work at Enterprise Speed
From AI Policy to Operating Control
How Leaders Make Governance Work at Enterprise Speed
Principles matter, but teams need decisions they can apply under delivery pressure. This session shows how to translate AI policies and regulatory frameworks into ownership, decision rights, evidence, escalation paths, and review points. Vivek connects board-level intent to the work of product, engineering, privacy, security, legal, and procurement teams, helping leaders build governance that moves with the system rather than arriving after deployment.
Talk 03
When AI Can Act
Human Accountability in the Agentic Enterprise
When AI Can Act
Human Accountability in the Agentic Enterprise
The governance question changes when AI moves from recommending to using tools, delegating work, and taking action. Vivek prepares leaders for that shift by examining authority limits, human checkpoints, exception handling, audit evidence, and the conditions under which an agent should stop. The session gives organizations a practical vocabulary for preserving human accountability without freezing responsible experimentation.
03 / Original framework
Governance for the layer most assessments miss.
PAARA is Vivek's practitioner methodology for evaluating connector-layer risk in generative and agentic AI.
PAARA
Privacy-Aware AI Risk Architecture
PAARA examines what enterprise AI can access at query time, where its retrieval boundaries sit, how permissions accumulate, and what evidence exists when systems combine data or take action.
It complements established privacy, security, and AI-risk processes by adding a structured view of the connector layer.
Read the working paperAConnector inventory
BImpact triggers
CPermission risk
DRetrieval boundaries
EInjection exposure
FInsider amplification
GControl alignment
Applied research filings
Eight U.S. provisional patent applications filed as sole inventor, covering connector capability registries, automated connector review, control-decay monitoring, and AI risk evidence — the applied engineering work behind the PAARA method.
Provisional applications only. No issued patent is claimed or implied.
04 / About Vivek
Technical depth, translated into executive action.
Vivek Kumar helps executives answer a question most AI programs postpone: what can the system reach, combine, and act on once it is connected to the enterprise?
He is an AI governance architect, privacy and cybersecurity leader, and creator of PAARA, a practitioner methodology for evaluating connector-layer risk in generative and agentic AI.
His background spans enterprise responsible AI, data protection, information governance, and regulatory compliance across technology, analytics, healthcare, and insurance environments. Vivek has served in senior governance and data protection roles, contributed practitioner analysis to the IAPP, and spoken to executive, policy, and professional audiences.
He is an AI2030 Global Fellow and a researcher affiliated with Indiana Wesleyan University. His peer-review service spans the NeurIPS ethics review process, the ACM/AAAI Conference on AI, Ethics, and Society, and journals published by Elsevier and Springer.
His work gives leaders a practical way to connect AI ambition with access controls, evidence, accountability, and human oversight.
AIGP · FIP · CIPP/US · CIPP/E · CIPM · CISM
05 / Recognition
Selected by peers, not by application fee.
Merit-based recognition from academic committees and global AI leadership programmes.
Fellowship · Global
AI2030 Global Fellow
Merit-selected fellowship for practitioners advancing responsible and trustworthy AI, admitted through committee review.
Recognition · 2024
FDPPI Privacy Awards
Recognised by the Foundation of Data Protection Professionals in India for contribution to data protection practice.
Credentials
FIP · AIGP · CIPP/US · CIPP/E · CIPM · CISM
Fellow of Information Privacy, the IAPP's highest designation, held alongside certification across AI governance, privacy management and information security.
06 / Peer review and editorial service
Trusted to judge the field's own work.
Invited review and programme-committee service for leading AI, ethics and governance venues.
NeurIPS 2026
Conference on Neural Information Processing Systems
Ethics reviewer
AIES 2026
ACM/AAAI Conference on AI, Ethics, and Society
Programme committee
Journal of Responsible Technology
Elsevier
Reviewer
AI and Ethics
Springer
Reviewer
Global South AI Safety Hackathon
Apart Research · AI Governance and Policy track
Judge
IEEE and Springer conferences
FTNCT, IATMSI, ICiTsIF, GAISS and SASIGD programme committees
Technical reviewer
07 / Public work
Published, heard, and in the room.
Selected public work that shows the range of Vivek's contribution: analysis, research, conversation, and practitioner review.
IAPP analysis
Privacy programs can't see AI connectors
Connector-driven data flows and the emerging insider-risk blind spot · 2026
Working paper
PAARA: A Connector-Layer Risk Assessment Methodology
SSRN working paper · DOI 10.2139/ssrn.6784078
Peer-reviewed
IEEE SoutheastCon 2026
Conference paper · DOI 10.1109/SEASCON59249.2026
Book chapter
Ethical Consideration and Challenges in Quantum AI
De Gruyter · DOI 10.1515/9783112217269-013
Podcast
The Data Diva Talks Privacy
Conversation with Debbie Reynolds on privacy-aware AI risk · 2023
Industry panel
Practical AI: Use Cases That Deliver Value
Info-Tech LIVE, Las Vegas · 2024
Bring sharper AI governance into the room.
For keynotes, executive briefings, panels, podcasts, and professional or university programs.
Start a conversation