Models, banking applications, product economics and AI governance.
OpenAI introduced GPT-6 Sol and Luna with lower API prices. Per million tokens, Sol costs $2 for input and $10 for output; Luna costs $0.10 and $0.50. The company reports improvements in professional work, coding and computer use.
Published: Sep 27, 2026Google described an extension of Private AI Compute designed to retain assistant context across sessions and devices. Its proposed architecture combines encrypted storage, keys held by user devices and processing inside secure enclaves.
Published: Sep 27, 2026Utah 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.
Published: Sep 27, 2026Anthropic announced an embedded evaluation partnership with Accenture, led by its AI business Faculty, giving evaluators deeper access during model development. The program covers adversarial testing, alignment and safeguards. Anthropic and Accenture each expect to invest at least $1 billion in evaluation capacity over five years.
Published: Sep 27, 2026An APort-authored preprint replayed 4,371 human-written attacks across 14 models. In 68,970 matched tests at policy levels 2–4, it reports 105 transfers to prohibited recipients with model-only controls and none when a deterministic authorization check guarded tool execution.
Published: Sep 27, 2026Affirm 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.
Published: Sep 27, 2026The Utah Bankers Association lists the AI-Native Banking and Fintech Conference for September 29, 9 AM–6 PM Mountain Time, at the Ken Garff University Club at Rice-Eccles Stadium.
Published: Sep 27, 2026Bank warnings about AI shopping bots bring scams, privacy, steering and customer recourse into focus. Agent adoption creates a new question: can an issuer reconstruct what the customer actually authorized?
Published: Sep 26, 2026Anthropic’s Claude Opus 5.5 capability, cost and safety claims were reported September 22. Vendor benchmarks need to be assessed separately from production evidence.
Published: Sep 26, 2026A practical evaluation of Fraud Signal, custom-model hosting and agent-assisted investigations, with separate evidence standards for prediction and workflow automation.
Published: Sep 27, 2026How behavioral intelligence can complement identity and transaction controls, why unusual behavior is not proof of fraud, and how to evaluate newer sequence-model research.
Published: Sep 27, 2026An evidence-focused review of device signals, issuing-risk models and foundation-model claims, including how to interpret AUC-PR and test cross-institution performance.
Published: Sep 27, 2026What document-authenticity models can detect, how their verdicts differ from verified income, and how to handle benign edits, forged statements and customer review paths.
Published: Sep 27, 2026The April 17, 2026 interagency guidance supersedes SR 11-7 and SR 21-8, emphasizes materiality and excludes generative and agentic AI from its formal scope without removing broader governance responsibilities.
Published: Sep 27, 2026What Taktile publicly describes, what the evidence does not establish, and how a bank can test decision quality before scaling automation.
Published: Sep 27, 2026A practical review of Oscilar’s rules, machine-learning and agent capabilities, with tests for alert quality, human approval and data protection.
Published: Sep 27, 2026What SentiLink’s technical materials say about scores and explanations, and how to evaluate them without treating a risk rank as a probability or a legal conclusion.
Published: Sep 27, 2026A framework for evaluating Provenir’s AI capabilities, explainability claims and operating economics across the credit lifecycle.
Published: Sep 27, 2026A proprietary lending system case study: architecture, conversion evidence, incremental credit economics, explainability and the limits of public validation.
Published: Sep 27, 2026A practical assessment of AI underwriting, the evidence behind vendor claims and the controls needed for a credible pilot.
Published: Sep 26, 2026Understand the score, validate the intervention and distinguish synthetic-identity risk from ordinary credit risk.
Published: Sep 26, 2026Fraud decisioning, digital trust and the evidence needed to evaluate performance claims.
Published: Sep 26, 2026Updates the structure and references of the Cybersecurity Supervision Work Program and rescinds the 2023 bulletin. The OCC says the update adds no procedures and establishes no new regulatory expectations.
Source date: Sep 21, 2026Replaces SR 11-7 and SR 21-8; emphasizes a tailored, risk-based approach.
Source date: Apr 17, 2026CFPB status page notes the October 29, 2025 stay of compliance dates.
Source date: Oct 29, 2025Reconsideration questions on representatives, fees, security and privacy.
Source date: Aug 22, 2025Final guidance on identity proofing, authentication and federation, replacing SP 800-63-3. Useful for evaluating assurance levels and vendor controls; it does not independently establish compliance with bank CIP requirements.
Source date: Jul 31, 2025Open-banking provisions; read with the compliance-status page and pending reconsideration.
Source date: Oct 22, 2024Generative-AI risk considerations and suggested actions.
Source date: Jul 26, 2024AI governance and risk-management framework; not a banking regulation.
Source date: Jan 26, 2023Privacy notices and limits on disclosure of nonpublic personal information.
Information-security requirements for financial institutions within FTC jurisdiction.
OCC notification requirements for covered incidents and bank service providers.
Reasonable measures for disposing of consumer-report information and records held for a business purpose. Relevant to retention schedules, document destruction and disposal service-provider oversight.
Statutory recognition of electronic signatures and records, with conditions for electronic delivery of legally required consumer disclosures. Useful for consent, retention and reproducibility controls in digital account opening.
Official explanation of consumer data rights, covered-business thresholds and privacy responsibilities, with a link to the Act. Check statutory exclusions and financial-services exemptions before applying the general requirements.
Official access point for Utah AI policy, regulatory-relief work and the supervised AI Sandbox. Program participation and proposed policy changes should not be treated as general exemptions from financial-services law.
Official regulation, coverage guidance, exemptions, incident reporting and compliance resources for DFS-covered entities. State licensing and the applicable exemption determine coverage; a federal bank charter alone is not the test.
Official framework, implementation resources and profiles for organizing cybersecurity risk management. A voluntary framework for most private-sector users; distinguish its outcomes from binding banking requirements.
A customizable security and privacy control catalog. The official page links Release 5.2.0, issued August 27, 2025, and its changes. Useful for mapping controls; adopting the catalog is not evidence that controls operate effectively.