AI Sovereignty Versus Internet GovernanceThe 2026 eCrime Measurement, Risk and Policy Roundtables will examine how Internet governance is evolving. Specifically, this year’s colloquium will consider how hyperscaled vertically integrated computational infrastructure providers and AI reasoning platforms have evolved into sovereignties in their own right. If you’re coming to this years policy roundtables, please register now here: https://apwg.org/events/ecrime2026#roundtables2026 Within this dialog, delegates will examine how these new and still evolving Internet infrastructure titans are incrementally displacing contemporary Internet-governance frameworks as currently curated by groups like ICANN and ISOC and IGF/UN coordination. Day's agenda follows: 9:30-9:45: Introductions and welcome
9:45-10:15 Mapping Infrastructure Authority Migration 10:15-10:45 Public Governance Authority at Risk
10:45-11:00 Break
11:00-11:30 Evidence, Cybercrime, and Lawful-access Continuity
11:30-12:00 Semantic Authority and Domain-Specific Ontologies
12:00-13:00 Lunch 13:00-13:30 EU Regulatory Leverage and Public-interest Compulsion
13:30-14:00 Corporate AI sovereignty compact 14:00-14:30 Coffee, Bites and Wrap up ROUNDTABLES PROGRAM DISCUSSION: Just a quarter century ago, the Internet's institutional architecture was programmatically distributed among constituencies representing different technical operations, technical standards, commercial interests and governance perspectives, embodied by organizations such as: ICANN; Registries and Registrars; IETF; RIRs; ISPs; and telcos. This multi-stakeholder model was itself, in part, a rejection of the national sovereignty/licensing model framed by the ITU and its suzerainty over telecommunications operating frameworks like SS7. In 2026, today’s Internet governance model is being challenged by the operational
hegemony of integrated digital infrastructure platforms (e.g. AWS, Microsoft, Alphabet/Google, Oracle); Internet intermediation services providers (Cloudflare, Akamai); and AI reasoning platforms (e.g. OpenAI, Anthropic, etc.) as these hyperscaled Internet infrastructure powers exercise their own separate sovereignty over compute infrastructure, network mediation, AI reasoning, and enterprise operations, becoming constitutional actors in their own right. The 2026 eCrime policy roundtable session will ask: Does this shift in the powers commanding authority over so many data logistics and archival functions, including content, communications, compute, identity, personal information, code, (end-user) devices, and so on require a civilization-level response? Do these transformations of anti-cybercrime enforcement power, evidentiary authority, and operational data-access necessitate prescriptive domain-specific (e.g. DNS; SS7; AI guardrails) ameliorative policy engagements? And, if so, what would those entail—and which political and diplomatic venues would develop and curate those updated policies? The eCrime symposium has been holding our policy roundtable every year at the conclusion of eCrime since 2024. Our first year was hosted at the Fletcher School of Government and Diplomacy at Tufts University. Last year, CAIDA at San Diego Supercomputer Center was our host. These eCrime Measurement, Risk and Policy Roundtables are APWG eCrime’s attempt to measure the gap between cybercrime’s advances and the policy infrastructure (public and private sectors both) that is required to manage contemporary cyber threats like other kinds of persistent, civilization wide risks such as those related to public health, underwriting risks, etc. This link provides full draft prospectus in PDF format, for viewing and printing: https://drive.google.com/file/d/1Sathxm1skFyj0OczZbm36mUedxzS237r/view | 
