The AI Risk & Compliance Network presents
AI Oversight
Summit 2027
October 2027Two days
Kyoto, JapanEnglish and Japanese
First edition, Kyoto
京都 二〇二七年十月
About the summit
AI has started acting, not just advising. Oversight has not caught up, and someone has to sign their name to it. For two days in Kyoto, the people accountable for AI at regulated financial institutions meet regulators, builders and each other to work out what accountable oversight of AI agents looks like in practice.
Why Kyoto, why now
Firms are moving from models that recommend to agents that execute: payments, trades, claims, customer decisions. Every control framework in use today assumes a human reviews the output.
- 01
Agents
What does human oversight mean when no human reviews each action? Every compliance officer with an agent in production is now asking that question, and supervisors are starting to ask it back.
- 02
The rules
Rules for AI in finance are arriving across Europe and Asia, and each jurisdiction is taking its own line. Firms that operate across borders need one answer for several regimes.
- 03
Kyoto
Japan has played a central part in international work on AI governance, and Kyoto is where the world has signed things before. The summit closes with a short statement of its own.
Built for the person who signs off
Chief compliance and risk officers, general counsel, heads of model risk, AI governance officers, and the board members who oversee them, from banks, insurers, brokers, asset managers and exchanges. Vendors are welcome as sponsors and exhibitors. The program is written for the people who answer to the regulator.
| Most AI governance events | This summit |
|---|---|
| Broad responsible AI, ethics and policy | One job: the person accountable for AI at a regulated firm |
| EU and US regimes at the centre | Japan, Korea, Hong Kong and Singapore in one room, with the EU and US as counterparts |
| Vendor-led content | Practitioner content; vendors on the Expo Stage |
| Ends with slides | Ends with the Kyoto Statement |
Program at a glanceプログラム
In development. Working session titles.
Day 0
Members, the day before- Network roundtable (closed, Chatham House Rule)
- First look at the member benchmark
- Drafting the Kyoto Statement
- Dinner for speakers and partners
Day 1
What oversight means now- What supervisors will ask next
- Human oversight when the human isn't in the loop
- AI incidents: what to report, to whom, how fast
- One answer for several regimes
- Case clinics: model risk for LLMs; an AI incident, start to finish
- Garden reception
Day 2
Who is accountable- What the board needs to see
- Liability: who pays when the agent is wrong
- Selling into Europe
- Case clinics: building the AI inventory; AI in the three lines of defence
- The Kyoto Statement on AI Oversight
Formats
The Kyoto Statement on AI Oversight京都声明
A short statement on AI oversight, written by the people who do it. Network members draft it on Day 0. Signatory institutions present it on the Main Stage on Day 2. It sets out what accountable oversight of AI agents should look like at a regulated institution, in plain language a board can adopt.
Review or sign the draft →Six groups, one room
- 01Regulators and supervisors
- 02Banks, insurers, brokers and exchanges
- 03AI labs and agent builders
- 04Cybersecurity and AI red teams
- 05Law firms
- 06Standard setters and associations
Regulators speak at the summit. They are never sponsors.
Call for talks
We are looking for practitioners with a real case: a control you built, an incident you handled, an examination you went through. Practitioners get priority over product talks.
Propose a talk →Get involved参加方法
Registration opens in 2027. Write now and you will hear first.