Snippings

xxx

Sullivan & Cromwell told a US federal bankruptcy court that a major filing it made in a high-profile case contained multiple “hallucinations” made by AI software.

Andrew Dietderich, the head of S&C’s restructuring practice, apologised in a letter to New York federal judge Martin Glenn on Saturday for mistakes that included misquoting the US bankruptcy code and citing cases incorrectly in a court filing made on April 9.

“We deeply regret that this has occurred,” he said in the letter.

Dietderich said the firm’s policies on the use of AI had not been followed when the document was prepared, and it was considering whether it needed to make “further enhancements” to its internal training and review processes.

From: Elite law firm Sullivan & Cromwell admits to AI ‘hallucinations’.

xxx

xxx

ARTIFICIAL INTELLIGENCE
BankThink The AI prompt that could wipe out banks’ net interest income
By Michael Abbott
Published February 03, 2026, 8:58 a.m. EST
4 Min Read
Facebook
Twitter
LinkedIn
Email

It won’t be long before bank customers can ask an AI agent to optimize their returns on idle cash. When it happens, banks’ net interest income is going to come under direct threat, warns Michael Abbott, of Accenture.Adobe Stock
When JPMorganChase CEO Jamie Dimon speaks, the industry listens. His comments on competition at this year’s World Economic Forum were telling.

Fintechs aren’t letting up: “They’re coming at you, some to pick a sliver of the business, some to take your whole business.”

AI could restructure customer access and product delivery. “Agents … may lead to us winning and losing big areas.”

Banks could face significant losses if they fail to adapt. “If you put your head in the sand, you will lose.”

The future of competition for banks will be fierce in 2026. However, I believe a new class of competitors will emerge that attack banking from the outside-in and target the balance sheet.

The last 25 years of fintech disruption have seen fintechs attempt to mimic banks and compete for the same services. They’ve seen limited gains: After 25 years of digital not a single neobank has cracked the global top 200 banks by assets.

Drive auto financing forward with instant payments
With instant auto loans, you can improve the car-purchasing experience for buyers and sellers and create new loan opportunities for your institution.
PARTNER INSIGHTS FROM FEDERAL RESERVE FINANCIAL SERVICES
With the emergence of AI engines, we see a very different competitive landscape emerging, with tech companies that attack both sides of the balance sheet but never touch banking itself. They will be infinitely more disruptive than fintechs already operating within the system.

It won’t be long before customers can ask their favorite GPT engine this simple question: “Help me optimize the return for my idle cash.” That one prompt takes direct aim at deposit beta in the U.S banking system.

Many customers in the U.S. are ready for it: Seventy-one percent say they would use an “everywhere” AI financial assistant that helps manage all their banking relationships, and 55% would trust a generative AI platform to own that experience, according to Accenture’s recent Future of Banking Experience survey.

Even a relatively small shift in pricing could lead to a 5% drop in lending margin and 15% drop in deposit margin, putting around 22% of U.S. banking pretax income at risk, according to Accenture Research.

From: The AI prompt that could wipe out banks’ net interest income | American Banker.

xxx

xxx

Rep. Bill Foster, D-Ill., highlighted the fragmentation of identity standards across jurisdictions and the operational challenges that fragmentation creates for financial institutions. In a setting increasingly shaped by automated agents and machine-driven transactions, he observed, banks must be able to determine “who is the legally traceable human behind this agent that has just contacted my agent.

From: Treasury Chief Says Smarter AML Starts With Better Identity.

xxx

xxx

A test of IATA’s face biometrics-based digital identity for air travel for a journey beginning with Japan Airlines (JAL) at Tokyo’s Haneda Airport was the first to involve a flight transfer and different digital wallets.
The airline shared details of the trial, for which it worked with Tokyo International Air Terminal Corporation (TIAT). The trial tested IATA’s proof-of-concept (PoC) for Contactless Travel at Scale, which aims to enable fully contactless, biometric‑enabled international travel through its One ID initiative. Hong Kong International Airport, London’s Heathrow Airport and British Airways also participated in the trial.
The service was tested by travelers taking off from Tokyo Haneda Airport and transferring in Hong Kong for a flight to the UK.

From: IATA digital ID trial shows interoperability across countries, wallets and biometrics | Biometric Update.

xxx

xxx

Two new papers possibly bring Q-Day closer, and perhaps show that the 2035 target for the switch-over to quantum-robust cryptography is perhaps under review. Each of the papers outlines attacks on elliptic curve cryptography (ECC), and which is commonly used for digital signatures (ECDSA and EdDSA) and key exchange (ECDH). This includes its application in blockchain, TLS sessions, SSH, and many other secure protocols.

From: Classical Public Key Methods Are Crumbling Under The Threat of Quantum Computers? | by Prof Bill Buchanan OBE FRSE | ASecuritySite: When Bob Met Alice | Apr, 2026 | Medium.

xxx

xxx

Hours after Anthropic said it was investigating a report that users had gained unauthorized access to its ballyhooed Mythos model, OpenAI accidentally made a slate of its own unreleased models available on its Codex app.

Anthropic has gone to great lengths in recent weeks to stress that Mythos is capable of devastating cyberattacks, and that it was only making the model available to a select few companies and government agencies as a result.

Even so, a group of unauthorized users had been accessing Mythos without Anthropic’s permission and sharing their findings in a Discord channel, Bloomberg reported Tuesday

From: New Security Breaches at Anthropic and OpenAI Proved Mark Zuckerberg Right — The Information.

xxx

xxx

American culture, which has grown more deliriously in favor of violence for years, has finally arrived at the point where that word, “violence,” is no longer sufficient. What we’re really talking about, specifically, is assassination, as the figure of the “‘good guy”’ killer is now very much ascendant, and our country, shaped by the psychotic contours of the internet, is now very much inside a new assassination culture.

From: Assassination Culture.

xxx

xxx

World ID allows AI agents to be linked to verified human users. This setup is designed to support controlled automation in areas such as digital commerce and workflow systems.

From: World ID upgrade expands “proof of human” system across digital platforms | Bitget News.

xxx

xxx

Additional features include Concert Kit, which enables artists to allocate tickets specifically to verified humans, aiming to curb automated scalping. With its network now spanning 18 million verified individuals globally, the update arrives as demand grows for reliable human verification in an AI-driven internet.

From: World ID 2.0: Sam Altman’s Digital Identity Play Expands Beyond the Orb.

xxx

xxx

The founders who architect AI as a co-founder will have a structural advantage over those who treat it as a fancy autocomplete.

From: The One-Person Unicorn 🦄 – by Linas Beliūnas.

xxx

Carta data shows that solo founders went from a fifth of all startups in 2019 to a third in 2025. My guess is that it will be more than half in 2030.