AI & Finance News — 2026-09-19 (Sat) Morning News

Overview

Implementation of financial agent-type AI (agentic AI) is progressing simultaneously in banking, payment, and underwriting operations, and state regulatory authorities are also rushing to develop supervisory frameworks. Meanwhile, the enthusiasm for AI-related trading is paralleled by regulatory and fraud risks (PYMNTS.com, CNBC, Consumer Finance Monitor).

Bank trends

Goldman Sachs said consumer agent AI represents a new phase of growth (Goldman Sachs), and John Flood said non-AI results were also "very strong" (CNBC). M&T Bank expects AI and fee business to contribute to loan growth in 2026 (Yahoo Finance). In Japan, Hokkoku Bank and Hokuriku Bank are using AI to speed up corporate loan screening and develop auxiliary tools (Hokkoku Shimbun). On the other hand, there have been reports of Android malware exploiting AI targeting bank login information (malwarebytes.com). CSBS releases AI oversight playbook for state inspectors, highlighting oversight gaps at the federal level (PYMNTS.com).

Insurance trends

A bill has been introduced in New Jersey that would target the denial of insurance claims due to AI (Insurance Business). GenWay brings AI automation to non-QM underwriting (nationalmortgageprofessional.com). PYMNTS.com asks whether the enterprise insurance market could be the "hidden ceiling" for agent-based AI adoption. When introducing AI underwriting, Affirm CEO said, ``We have not cut staff to purchase AI tokens.'' (CNBC) While the introduction of legal AI is facing difficulties, a success story of a global insurance company was also reported (JD Supra).

Trends in asset management and market transactions

Baillie Gifformd and Essentia Analytica partner on AI program (Markets Media). While the use of AI in healthcare strategies is progressing in private equity (Modern Healthcare), Coforge pointed out the widening gap between expectations for AI and actual returns (ET Enterprise AI). Open source library of AI agents for commercial real estate expands to 90 types (businessinsider.com). In the market, bullish option trading of AI-related stocks has rapidly increased (CNBC, qz.com), and it has been pointed out that Nvidia's "AI token trading" is "correlated in attention but not returns" (Benzinga). There was also an incident where 274.6 ETH was stolen from 224 people due to a fraudulent AI trading bot tutorial (Crypto News).

Fintech/payment trends

Stripe is incorporating AI bots into wallets (Payments Dive), and Mastercard and Visa are looking to introduce AI bots to checkout, but consumers still want to make the final click themselves (Fortune). Ripple will reportedly connect XRP to Stripe's AI agent payments, and payment sessions will run in XRP instead of RLUSD (Yahoo Finance). Mastercard grants spending management functions to AI agents through Alchemy (TradingView) amid flat stock price. In Japan, a new technology was reported that uses AI to autonomously perform payments by comparing provenance and on-chain payment history to confirm authenticity (tsuhan-ec.jp).

Remarks from key figures and authorities in the financial world

- Goldman Sachs John Flood (executive): said that the financial results were "very strong" in areas other than AI (CNBC). - Affirm CEO: Says the company has never cut staff to purchase AI tokens and is expanding its AI underwriting (CNBC). - Luca Ferrari (related to Bending Spoons, title not listed): Airtable, AI agents, and why Bending Spoons is not private equity (finance.biggo.com). *There are no statements or references in this headline regarding the key figures being tracked (Jamie Dimon, Larry Fink, Jane Fraser, David Solomon, Brian Moynihan, Ted Pick, Ken Griffin, Paul Atkins).

Trends in AI-related stocks and funding

Chinese AI startup Manus aims to raise $500 million and is considering a Hong Kong IPO (WSJ, Seeking Alpha). This is said to be a move after the separation of Meta. Predictive AI startup Mantic raises $25 million (Reuters). Regarding Nvidia stock, the tone is asking whether it is ``enthusiastic buying'' or an ``AI bubble'' (Yahoo Finance UK). It was recalled that the CEO of OpenAI and Anthropic once warned of the "risk of AI extinction" (Investor's Business Daily).

Trends in financial regulation and supervision

The Conference of State Bank Supervisors (CSBS) published the Artificial Intelligence Supervisory Framework for state examiners to guide state reviews of financial AI risks (Consumer Finance Monitor, Retail Banker International). There are still gaps in the supervisory framework at the federal level (PYMNTS.com). South Africa has put AI regulations on hold pending the opinion of global regulators (Bloomberg). There was also an argument that AI governance requires a macroprudential perspective (overlooking the entire financial system) (promarket.org).

Integrated analysis: What is happening now

What is common in the banking, insurance, and payment sectors is that agent-type AI is starting to enter the center of decision-making in ``screening, underwriting, and payment execution.'' Hokkoku Bank's loan review, GenWay's non-QM underwriting, and Stripe/Mastercard/Visa's autonomous payment bots are all moving toward reducing the final human confirmation process. In response, it has become clear that state regulatory authorities (CSBS) are taking the lead in developing supervisory frameworks and filling the void at the federal level, but there are also moves, such as in South Africa, to wait for guidelines from international supervisory bodies, and the timing of regulations varies by country and region. In terms of capital flows, the market remains optimistic, with large-scale procurements continuing for specialized AI startups like Manus and Mantic, and bullish options piling up on "AI token trading" centered on Nvidia, but as Coforge points out, the gap between expectations for AI and actual returns is widening in the PE industry, and a gap is emerging between overheating and penetration into the real economy. As for risks, fraud by malware and fake AI trading bots that exploit AI are becoming more apparent among consumers and individual investors, and the benefits of introducing AI (improved efficiency and speeding up examinations) are concurrently progressing with the expansion of the targets of attacks that exploit AI.

Future points of interest

- How will the CSBS AI supervisory framework be applied to actual inspections, and how will the federal response (FRB, OCC, FDIC, etc.) follow suit? - Manus's progress in raising $500 million and considering an IPO in Hong Kong, and market reaction regarding valuation. - Will state-level regulations regarding insurance claim AI rejections (such as the New Jersey bill) spread to other states?