Liability looms over AI shopping
Banks and payment firms raise concerns over missing legal and liability frameworks for agentic commerce applications.
Payments Dive·28 Oct 2026
Use this view to inspect the underlying evidence corpus. For ranked developments, decision posture and interpretation, use Signals.
Raw feed or Signals?
Raw feed is chronological evidence. Signals ranks and interprets material change.
Banks and payment firms raise concerns over missing legal and liability frameworks for agentic commerce applications.
Banks risk policy non-compliance as unofficial AI usage expands across credit scoring, document analysis, and loan memo drafting.
Oracle will natively offer Google Tensor Processing Units to cloud customers, marking the first rival hyperscaler integration for TPUs.
UK policy will prioritize grid access and on-site power generation for critical data centers to accelerate infrastructure development.
Hummingbird acquired AML machine learning platform Quantimon to integrate risk detection into its compliance orchestration suite.
Bloomberg introduced BQuant Assistant, an AI coding assistant integrated into BQuant Desktop and Enterprise for quantitative analysis.
Casap raised $56M in Series B funding led by Base10 Partners with participation from Experian and Synchrony for its AI disputes platform.
Zocks launched an AI feature automating account opening workflows during client onboarding for financial services firms.
Tech firms face an influx of software bug reports as automated AI vulnerability scanners flood development teams with flaws.
Ant International CEO Peng Yang argued at UNCTAD that AI connectivity will drive re-globalisation by expanding SME market access.
OpenAI and ServiceNow expanded their partnership to integrate OpenAI models and agents into ServiceNow enterprise workflows.
NatWest deployed AI agents from Adclear to automate marketing compliance approvals, reporting an 80% reduction in review times.
Hummingbird acquired machine learning platform Quantimon to integrate AML risk detection and federated learning into its compliance workflow platform.
Zafin has agreed to acquire Naehas to expand its AI-enabled banking platform with offer management and disclosure automation capabilities.
Lloyds expanded an agentic AI platform to support fraud specialists after testing across 10,600 live debit card fraud cases.
LuLu Financial Holdings and Purpleplum launched a UAE SME banking platform featuring agentic AI onboarding and digital accounts.
Anthropic expanded its AI for Science program, partnering with 20 global scientific foundations and research groups.
LangChain published technical guidance on using Openwiki knowledge graphs to structure long-term memory for autonomous AI agents.
NatWest partners with Adclear to deploy AI agents that reduce marketing compliance review times by 80% amid increased FCA scrutiny.
Microsoft and Alphabet earnings reflect growing investor scrutiny over heavy capital expenditure for AI infrastructure and profitability.
FCA-regulated prepaid account provider DOSH integrates Google Pay to extend contactless payment access to underbanked consumers.
An industry commentary piece on PYMNTS explores the critical role of permissioning and authorization frameworks for autonomous AI agents.
Sir Geoffrey Vos argues AI could adjudicate civil cases, raising questions about where to draw the boundary.
Advances in AI mathematical reasoning raise security concerns regarding classical global encryption, challenging long-term preparations.
Anthropic founders are reportedly tightening their internal voting control ahead of a potential blockbuster public offering.
Researchers propose PAI-KNN, an inference-time prior adjustment method for k-nearest neighbor classification under concept drift and imbalance.
An arXiv preprint evaluates temporal splits versus random splits for click-through rate model recalibration using Avazu data.
Researchers propose a model-agnostic method to quantify predictability in multivariable time series independently of specific models.
Researchers propose query-vision interaction decomposition for efficient visual token pruning in large vision-language models.
Researchers show adversaries can create LLMs that pass safety checks pre-transformation, but become malicious after quantization or pruning.