Search

Affirm reports more completed purchases from a hybrid underwriting model

Affirm announced a transformer-based underwriting model on September 17 and reported 3.4% more completed purchases against a control group. Its technical account describes a transformer feeding learned credit representations into an XGBoost risk model.

Analysis

…headline model metric. The new deep dive separates predictive lift from the economics and governance of the decision.

Altman links OpenAI IPO timing to confidence in model safety

The Verge’s September 29 report of an on-record post-DevDay discussion says Sam Altman linked taking OpenAI public to being able to make confident statements about model safety. He did not give a firm IPO timetable.

Source

The Verge — Altman’s post-DevDay remarks https://www.theverge.com/ai-artificial-intelligence/1002505/sam-altman-openai-ipo-devday-ai-safety

Anthropic prospectus puts AI economics, safety risk and founder control in one frame

Reuters reported from a confidential IPO prospectus that Anthropic generated nearly $4.6 billion of 2025 revenue while recording an $8.06 billion operating loss, spent $7.33 billion on compute and infrastructure, and plans $518 billion of future cloud, computing and infrastructure obligations. The filing also says two customers supplied nearly a quarter of revenue, devotes about 80 pages to risk factors and gives a founder vehicle 50.1% voting power over key matters.

Analysis

…combination of large fixed commitments, safety disclosures and founder control makes governance part of product risk: leaders may prioritize mission or caution over near-term returns, while the release cadence needed to compete can pull the other way. For financial institutions buying frontier AI, the…

Anthropic says government actions could spill into customer and partner relationships

Reuters reported from Anthropic’s confidential IPO prospectus that government contracts represent less than 1% of annual revenue, yet the company warns that government attitudes and actions could still damage commercial customer, partner, employee and investor relationships. The filing cites a February federal stop-use order, a Defense Department supply-chain-risk designation and June export restrictions that were later lifted.

Analysis

…That is a substantive extension of the previously reported prospectus economics and governance story, but it remains a company risk-factor disclosure rather than evidence that customers have left or that material losses have occurred.

Anthropic says most of its $518 billion compute plan is non-cancelable

Reuters reported from Anthropic’s confidential IPO prospectus that about 80% of its $518 billion of planned infrastructure commitments is non-cancelable or payable regardless of use. The filing describes at least $111.1 billion for Google, $110 billion for Amazon and $31.4 billion for Microsoft over seven to 10 years, plus about $161.2 billion of largely non-cancelable Broadcom-related equipment leases.

Source

Earlier Credit Current coverage · prospectus economics and governance https://thecreditcurrent.com/news/2026-09-29-0600-et-anthropic-ipo-prospectus-economics-governance-risk

Bank of America adds AI intelligence tools for treasury teams

Bank of America announced AskGPS Intelligence Hub on September 28, an internal suite for Global Payments Solutions employees that combines client, account and relationship information for treasury reviews, account-structure diagrams and relationship insights. The capabilities are planned in phases; employees retain responsibility for decisions and client interactions.

Analysis

The announced system extends Bank of America’s internal AskGPS generative-AI application from institutional knowledge access toward analysis of client and treasury relationships. The bank says the earlier tool, introduced in 2025, supports nearly 3,000 employees. The newly announced suite includes Intelligent…

Curtis backs proposed protections for AI whistleblowers

Utah Senator John Curtis announced his support for the bipartisan AI Whistleblower Protection Act. His office says the proposal would protect covered disclosures involving federal-law violations, national security or public safety, prohibit retaliation and prevent nondisclosure agreements from blocking protected reports.

Analysis

Analysis: The proposal puts employee escalation and reporting channels on the AI-governance agenda. Organizations developing AI should compare existing reporting practices with the bill’s actual coverage and procedures.

Experian lets consumers connect credit information to Google Gemini

Experian announced on September 23 that U.S. consumers can authorize its app as a Google Gemini Connected App and ask conversational questions about their credit rating and key influencing factors. Existing members can connect an account, while new users can establish an Experian membership at no cost.

Analysis

The integration moves consumer credit information into a general-purpose AI interface, creating a new access channel rather than a new bureau score. Users authorize the connection and can manage it through Google’s Personal Intelligence settings. Experian says consumers will continue to access their…

Alloy: connected onboarding, fraud decisions and the cost of customer friction

How orchestration, fraud signals and AI assistance affect account opening and ongoing service, with separate evidence for each component.

Source

[2] Alloy, Actionable AI product documentation; reviewed September 27, 2026; vendor claims https://www.alloy.com/actionable-ai

BioCatch: behavioral intelligence, scam prevention and the customer experience

How behavioral context may support safer account and payment journeys, with separate evidence for deployed tools, early research and customer friction.

A behavioral language model is not a chatbot

…investigative visualization and a newer sequence-model concept may sit under the same AI narrative but have different maturity and validation evidence. Procurement should identify which version and function would actually be deployed, what inputs it requires and which outputs are contractually suppor…

ComplyAdvantage Mesh: screening, payment service and investigation workload

How financial-crime data and case workflows affect onboarding and payments, with separate tests for matching quality, completion time and operating effort.

Different checks answer different questions

…The vendor describes both algorithmic matching and rule-based functions, so a single AI label does not identify the work being performed. [1]

Cursor and Anysphere: AI coding after the SpaceX acquisition, with financial-sector controls in focus

Cursor is Anysphere’s AI software-development platform. Following its August 2026 acquisition by SpaceX, the relevant questions include model strategy, enterprise economics, cloud-agent data boundaries and whether faster engineering produces dependable financial software.

Revision summary

Cursor is Anysphere’s AI software-development platform. Following its August 2026 acquisition by SpaceX, the relevant questions include model strategy, enterprise economics, cloud-agent data boundaries and whether faster engineering produces dependable financial software.

Databricks: the data platform behind enterprise AI, and the controls that matter in finance

Databricks combines data engineering, analytics, governance and AI on a consumption-based platform. The platform’s substantial adoption does not establish lower risk; data boundaries, operating costs and model dependencies vary by workload.

Revision summary

Databricks combines data engineering, analytics, governance and AI on a consumption-based platform. The platform’s substantial adoption does not establish lower risk; data boundaries, operating costs and model dependencies vary by workload.

Featurespace: behavioral analytics, payment acceptance and the cost of intervention

Featurespace’s behavioral fraud analytics in a named NatWest deployment, with reported scam metrics, customer commentary and tests that distinguish detection value from realized loss prevention.

Operating trade-offs

Real-time risk scoring can reduce exposure but adds latency, integration and model-governance demands. A bank should test performance during peak volumes, vendor outage and degraded data quality; define a conservative fallback and review alert queues. Model outputs should not automatically close alerts…

Feedzai: safer payments, legitimate transactions and financial-crime operations

How a fraud platform fits into payment acceptance and investigations, with distinct measures for detection, customer friction and operating economics.

Evidence gates · Table row

Governance · Reason codes, versions, overrides, fairness and monitoring · Challenge, audit and controlled change

FICO Platform: customer decisions, operating capacity and financial value

Decision software connects data, models and business rules to customer outcomes. Examine FICO Platform through service completion, peak demand, integration costs and the limits of its reported lending examples.

Revision note

Broadened the profile from lending governance to customer journeys and operating economics; added a capacity-versus-cash-savings example, retained dated customer evidence and corrected the historical guidance citation.

NIST AI Risk Management Framework ↗

AI governance and risk-management framework; not a banking regulation.

Description

AI governance and risk-management framework; not a banking regulation.

AI Incident Database ↗

Documented AI incidents

Description

Documented AI incidents

Anthropic ↗

Model, product and research announcements

Arize AI ↗

AI evaluation and observability

Description

AI evaluation and observability

Credo AI ↗

AI governance tools

Description

AI governance tools

Fiddler AI ↗

Model monitoring and explainability

Google DeepMind ↗

AI research and models

Description

AI research and models

Hugging Face ↗

Model cards, datasets and technical resources

Meta AI ↗

AI models and research

Description

AI models and research

Alloy

Financial technology company · Fraud & identity, Fraud risks & controls, Customer onboarding, Decisioning & orchestration, Credit risk & analytics · 0 linked stories · 3 related research pages

Amazon Web Services

Technology company · Also known as AWS · AI infrastructure, AI security & agent controls, Enterprise technology · 1 linked stories · 4 related research pages

Description

Technology company · Also known as AWS · AI infrastructure, AI security & agent controls, Enterprise technology · 1 linked stories · 4 related research pages

American Fintech Council

Industry association · Financial technology, AI governance · 1 linked stories · 0 related research pages

Description

Industry association · Financial technology, AI governance · 1 linked stories · 0 related research pages

Bank of America

Bank · Banking, Payments, AI in financial services, Fair lending · 2 linked stories · 5 related research pages

Description

Bank · Banking, Payments, AI in financial services, Fair lending · 2 linked stories · 5 related research pages

CALM xAI

AI technology company · AI security & agent controls, AI governance · 1 linked stories · 0 related research pages

Description

AI technology company · AI security & agent controls, AI governance · 1 linked stories · 0 related research pages

Celtic Bank

Bank · Utah banking, Industrial banks, Small-business lending & finance, SBA lending, Bank & fintech partnerships · 0 linked stories · 6 related research pages

Citi

Bank · Also known as Citigroup · Banking, Payments, Digital assets, AI safety & evaluation · 3 linked stories · 4 related research pages

Description

Bank · Also known as Citigroup · Banking, Payments, Digital assets, AI safety & evaluation · 3 linked stories · 4 related research pages

Databricks

AI and enterprise technology company · Enterprise AI, Data governance, Financial technology · 0 linked stories · 1 related research pages

Description

AI and enterprise technology company · Enterprise AI, Data governance, Financial technology · 0 linked stories · 1 related research pages