MAS: AI Agents Could Move Deposits Faster, 3 Days After Barr

Singapore’s central bank warned on Thursday that AI agents could move deposits and other funds more quickly, raising new liquidity and financial-stability risks, Bloomberg reported.

“Agents have the potential to move money and allocate assets more quickly for smaller incentives”, Monetary Authority of Singapore managing director Chia Der Jiun said in his keynote at the Singapore FinTech Festival, according to Bloomberg.

Business Pill · Least Privilege

An evergreen Business Engineer explainer on least privilege and access control. It teaches the general idea only and does not discuss MAS or this speech.

The key insight: Singapore’s central bank says AI agents could move money faster for smaller incentives. It is the second central bank in three days, by our count, to warn about agent-driven deposit moves, after Fed Governor Barr on 16 November.

What Chia Said

Chia called on financial institutions to cap transaction instructions from AI agents to manage liquidity risk, according to a summary of Bloomberg’s report. He also urged regulators to watch the implications for financial stability.

He urged caution about people handing their financial account login credentials to AI agents, saying the safety of that approach has not been sufficiently assured, the same summary says.

“The future is likely to be a mix of a human economy and a new agent economy”, Chia said, as relayed from Bloomberg. “Autonomy must have boundaries, and actions and outcomes must carry accountability.”

How agent-driven money moves differ from human ones, as MAS described them in reports of Chia's keynote, and t
How agent-driven money moves differ from human ones, as MAS described them in reports of Chia’s keynote, and the safeguards MAS proposes. Concept diagram, not data.

What Was Posted on X

Polymarket relayed the warning from Singapore’s central bank that AI agents could move deposits out of banks faster, and called it the risk of an “agentic bank run”. The phrase does not appear in the parts of Bloomberg’s report and its summaries that we could read; this piece uses Bloomberg’s report of the MAS keynote as the source.

Why Speed Is the Risk

The concern is speed and small triggers. A second summary of Bloomberg’s report says MAS sees agents moving capital between institutions over minimal differences in returns or incentives.

It adds that agents can optimise and trade around the clock, faster than manual processes. If many agents react the same way to the same signal, synchronised flows could be amplified, MAS assesses, according to that summary.

Polymarket summed it up on X as the risk of an “agentic bank run”. The phrase does not appear in the parts of Bloomberg’s report and its summaries that we could read.

What MAS Plans to Do

At the centre is a proposed “Know Your Agent” protocol, the Bloomberg summary says. It would give banks and payment firms a common way to check an agent’s identity and what authority the user has given it.

MAS also plans to keep testing AI-enabled transactions, including agents that monitor liquidity and pay with tokenised deposits or stablecoins, the summary says.

Three Central-Bank Warnings in Four Days

Fed Governor Michael Barr said on 16 November that agents “may accelerate deposit outflows during stress periods”, as we reported on Wednesday. MAS spoke three days later, by our count.

On Thursday the ECB put euro area banks’ AI-related lending at around €75 billion, and UK regulators said AI lets attackers exploit flaws within hours or days.

Arizona’s attorney general opened a criminal probe of OpenAI this week over what she called the apparent infiltration of banks by rogue AI agents.

The Structural Read

As we read the reports, MAS frames the risk as speed plus small triggers: money that used to stay put because moving it was a chore may move whenever an agent sees a slightly better rate.

As we read it, the proposed fix is identity and authority, not a ban: a Know Your Agent check and limits on what an agent can instruct, while MAS keeps testing agent payments.

Chia Der Jiun, managing director, Monetary Authority of Singapore, as reported by Bloomberg

“Agents have the potential to move money and allocate assets more quickly for smaller incentives.”

Three Implications

FOR BANKS Chia called on financial institutions to cap transaction instructions from AI agents to manage liquidity risk, according to a summary of Bloomberg’s report.

FOR PAYMENT FIRMS A proposed Know Your Agent protocol would give firms a common way to check an agent’s identity and authority, the summary says.

FOR CUSTOMERS Chia urged caution about handing financial account login credentials to AI agents, the summary says.

What Is Not Established

We could not read the speech itself: the MAS website showed a maintenance page on Thursday afternoon. Bloomberg’s article is paywalled beyond its first two paragraphs.

The reports give no figure for how fast agent-driven outflows could be, and no level for any cap on agent instructions. MAS has not said when a Know Your Agent protocol would apply.

This Week In The AI Supercycle

AI Lost Its Central Banker, the Lab Oligopoly, and the Future of Open Source

This week’s issue: the Nvidia leadership change, read as the AI build-out losing its central banker: who now finances the GPU landlords, backstops the data centres and sponsors open models, and why two labs are becoming the bank.

Read the full issue →

Business Engineer Framework

The Map of AI

A Business Engineer map of 200+ AI companies across 9 layers.

Read the Map of AI →

The Bottom Line

Singapore’s central bank says AI agents could move money faster for smaller incentives, and wants limits on agent instructions plus a Know Your Agent check. It is the second central-bank warning on agent-driven deposit moves in three days, after the Fed’s Barr.

97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. Chia’s first quote and the headline warning are from Bloomberg’s report of 19 November 2026, of which we read the first two paragraphs. The other quotes and details are from two published summaries of Bloomberg’s report (Bloomingbit and TechFlow). The MAS speech text was not available to us. ‘Three days later’ is our count from 16 to 19 November. Not investment advice; for information only.

Sources: Bloomberg, 19 Nov 2026 · Bloomingbit, summary of Bloomberg’s report, 19 Nov 2026 · TechFlow, summary of Bloomberg’s report, 19 Nov 2026 · Polymarket on X, 19 Nov 2026

Scroll to Top
FourWeekMBA