HSBC wins Hong Kong stablecoin issuer licence
HSBC has been granted a stablecoin issuer licence by the Hong Kong Monetary Authority. The approval gives the bank permission to issue stablecoins under Hong Kong’s regulatory framework.
Finextra ↗Trade Surveillance
US Treasury proposes AML rules for stablecoin issuers
The U.S. Treasury’s FinCEN and OFAC have issued a joint proposed rule to implement parts of the GENIUS Act for stablecoin issuers. The proposal would add anti-money laundering and sanctions compliance requirements for the sector.
Finextra ↗Treasury warns banks on Anthropic AI model
Anthropic announced Claude Mythos Preview, a new general-purpose language model aimed at improving cybercrime detection. It also launched Claude Managed Agents, allowing users to deploy AI agents within its platform.
Finextra ↗US launches crypto cybersecurity sharing channel
The US government has launched a cybersecurity information-sharing initiative for the digital asset industry. The move is aimed at helping crypto firms share threat information and improve defenses against cyber risks.
Finextra ↗Synthetic identities drive 8% rise in global fraud
LexisNexis Risk Solutions says its latest Cybercrime Report found global fraud rates rose 8% over the past year. The increase was driven by attacks on gaming, gambling and ecommerce, along with synthetic identities and agentic bots posing as humans.
Finextra ↗Binance adds prediction markets feature
Binance has launched a new prediction markets feature in its cryptocurrency app. The update expands the platform’s product offering but does not indicate any regulatory action or enforcement issue.
Finextra ↗NextGen Nordics to spotlight AI and fraud readiness
NextGen Nordics will return for its ninth year in Stockholm on 28 April 2026, bringing together banks, regulators, technology firms and fintechs. The event will present data on the innovation divide, AI integration and fraud prevention readiness.
Finextra ↗PayPal adds payment links to Canva
PayPal is integrating payment links into Canva so creators can embed payment options in their designs. The move expands PayPal’s reach into creator tools and makes it easier for users to collect payments directly from visual content.
Finextra ↗Kyndryl launches AI workflow tool
Kyndryl announced a new AI-powered capability called Digital Twin for the Workplace. It is designed to predict and resolve technology issues before they disrupt workflows.
Finextra ↗OwlTing expands OwlPay with Visa Direct
OwlTing announced an expanded collaboration with Visa to support Visa Direct within its OwlPay payment infrastructure. The move is intended to broaden stablecoin-related payment offerings and improve cross-border payment capabilities.
Finextra ↗AI World
Florida AG probes OpenAI over alleged ChatGPT role
Florida’s attorney general has opened an investigation into OpenAI after reports that ChatGPT may have been used to help plan a shooting at Florida State University. The incident killed two people and injured five, and a victim’s family says it plans to sue OpenAI.
TechCrunch ↗Mercor faces lawsuits after data breach
Mercor, a $10 billion-valued startup, is dealing with the fallout from a recent data breach. The company is now facing lawsuits and reportedly losing major customers.
TechCrunch ↗Control theory trims AI models during training
Researchers developed a technique that uses control theory to remove unnecessary complexity from AI models while they are still learning. The approach cuts compute costs without reducing performance, making training more efficient.
MIT News ↗Meta AI app surges after Muse Spark launch
Meta's AI app jumped from No. 57 to No. 5 on the App Store after the launch of its new Muse Spark model. The move suggests strong user interest in the updated product and a potential boost in adoption for Meta's AI offerings.
TechCrunch ↗Anthropic limits Mythos release over security risks
Anthropic said it restricted release of its new Mythos model because it can find security exploits in widely used software. The article questions whether the move reflects genuine cybersecurity caution or broader concerns about the model’s power and Anthropic’s interests.
TechCrunch ↗Google and Intel expand AI chip partnership
Google and Intel are deepening their AI infrastructure partnership to co-develop custom chips. The move comes as demand for CPUs rises amid a growing global shortage.
TechCrunch ↗Sierra pushes agent tools beyond click-based apps
Sierra launched Ghostwriter, an agent designed to build other agents. The company says its “agent as a service” approach could replace traditional click-based web apps with natural language instructions that create and deploy specialized agents.
TechCrunch ↗Meta unveils Muse Spark multimodal reasoning model
Meta Superintelligence Labs released Muse Spark, the first model in its Muse family. The model is natively multimodal and adds tool use, visual chain of thought, thought compression, and multi-agent orchestration.
MarkTechPost ↗MIT fellow advances ethics of new computing tech
Michal Masny, the NC Ethics of Technology Postdoctoral Fellow, is focused on dialogue, teaching, and research around the social and ethical dimensions of new computing technologies. The piece presents his work as part of MIT's broader effort to think through the impact of emerging tech.
MIT News ↗OpenAI adds $100 monthly ChatGPT Pro plan
OpenAI has introduced a new $100-per-month ChatGPT Pro subscription for power users. The plan fills the gap between the existing $20 and $200 monthly tiers.
TechCrunch ↗