The transition of the Apple Card from being financed by Goldman Sachs to JPMorgan Chase will probably cut out a smaller company, CoreCard, that currently handles the credit card. When JPMorgan takes over, it will likely drop CoreCard as the Apple Card’s processor in favor of its own in-house payment processing. Before it got the Apple account and its estimated 12 million users, CoreCard was considered a niche processor. CoreCard currently manages the day-to-day functioning of the Apple Card, ensuring that transactions are completed and handling the billing of users. It is also CoreCard that has been responsible for developing some of the distinctive features of the card alongside Apple, like its first-of-the-month billing cycle. CoreCard also developed the “payment wheel” graphic on its bills that show cardholders their projected interest costs, widely seen as consumer-friendly and educational. JPMorgan Chase will be responsible for the Apple Card once Goldman Sachs exits the consumer market. Because JPMorgan Chase has its own processing capabilities, CoreCard owner Richard Strange believes it likely that the Apple Card processing job will go in-house, dropping CoreCard. The transition will mean the loss of its biggest client.
Broadridge’s integration of Uptiq’s tech into its wealth management platform provides advisors access to agentic AI apps that surface the most relevant loan options, automating securities-based lending- (SBL) workflows, referral submission and covenant tracking
Broadridge Financial Solutions has announced a strategic partnership and minority investment in Uptiq, an AI platform for financial services. The partnership aims to modernize wealth management by addressing the growing demand for artificial intelligence in financial services and developing a better wealth lending process. Uptiq’s AI-powered tools and Broadridge’s Wealth Lending Network will enable advisors to deliver smarter lending recommendations, save time, and help clients access the liquidity needed to achieve their financial goals. The integration will streamline the process of accessing securities-based lending solutions, particularly for financial advisors and wealth management firms not affiliated with banks. Broadridge’s investment supports Uptiq’s growth and reinforces a shared vision for transforming wealth lending.
Meta’s new Brand Rights Protection features allow brands to report suspected scam ads at scale and even to request action on ads that don’t explicitly use their IP
Meta’s adding some new elements to its Brand Rights Protection system, which will give IP holders more capacity to report violations, as well as related concerns within their niche, helping Meta refine and improve its enforcement measures. First off, after beta testing it for a few months, Meta’s now going live with its expanded reporting functionality, which enables all businesses that are enrolled in Brand Rights Protection to report suspected scam ads at scale. And interestingly, Meta will also now enable brands to request action on ads that don’t explicitly use their IP, via a new “Other” category. As per Meta, brands can also use this to request action on: “Ads that promote products, services, schemes or offers using deceptive or misleading practices, including those suspected to scam people out of money or personal information, and otherwise do not involve a rights holder’s intellectual property.” Meta’s also redesigned its takedown request experience, with fewer steps required to lodge a concern, while it’s also added new search options within the “Reports” tab to make it easier to find info on your reports. The platform also now includes AI image matching, based on reference images of your products, to detect potential concerns across its apps, and as Meta looks to incorporate more shopping-focused elements, this is an important consideration, giving rights holders peace of mind in ensuring that they have some form of recourse to protect themselves.
Klarna to sell up to $26B in Pay in 4 US loans to Nelnet, enhancing funding flexibility ahead of IPO; structured forward-flow agreement offers predictable, off-balance-sheet funding
Klarna Group Plc agreed to sell as much as $26 billion of buy-now, pay-later loans to the student loan giant Nelnet Inc. as the fintech looks for ways to free up capital ahead of its public debut. The multi-year deal is structured as a so-called forward-flow agreement, where a buyer agrees to purchase loans before they have been originated. The agreement enables Klarna to sell newly originated, short-term, interest-free Pay in 4 receivables to Nelnet on a rolling basis. Over the life of the program, up to $26 billion in total payment volumes are expected to be sold. The transaction delivers scalable and efficient funding to power Klarna’s U.S. growth, while enhancing balance sheet flexibility and supporting long-term capital strategy. “This is a landmark transaction for Klarna in the U.S.” said Niclas Neglén, CFO at Klarna. “Our partnership with Nelnet allows us to scale a core product responsibly, while continuing to deliver smooth, interest-free payment experiences to millions of consumers.” The forward flow structure offers predictable, off-balance-sheet funding and underscores Klarna’s ability to structure and execute large-scale capital markets transactions. Klarna will continue to originate and service all receivables under the program, ensuring continuity and quality of experience for both consumers and merchant partners.
Google Messages is testing RCS’ new MLS encryption which makes E2E encryption possible across different RCS clients and providers
Google Messages is beginning to test the new Messaging Layer Security (MLS) protocol. Universal Profile 3.0 adds support for MLS, which makes E2E encryption possible across different RCS clients and providers. Google first announced its support for this interoperable protocol in 2023. The GSMA and Apple announced official adoption this March. Google Messages is now beginning to test MLS encryption for RCS. It starts with a new message “Details” (long-press on the chat/text) screen that’s fullscreen compared to the current approach. You get a preview of the message at the top, with Google also showing a “Status” section for “Sent” and Delivered” that explains the new checkmarks. We see Google using the latest single circle design that has yet to become widely available. There’s also a “From” section, while the bottom portion provides more technical details including Type, Priority, Message id and Encryption Protocol. This new design is not widely rolled out in the beta channel. It’s unclear if that’s also the case for MLS as the old UI makes no indication, while Apple has yet to specify when support is coming.
Embedding agentic AI into dispute workflows and fraud controls can help banks unlock RTP’s full potential by adding a predictive, self-learning layer that can autonomously detect, decide, and act
To thrive in a real-time payment’s world, banks must embed AI and cyber resilience into dispute workflows, fraud controls and compliance operations. To unlock the full potential of real-time payments (RTP), financial institutions worldwide are adopting intelligent, AI-led solutions to manage fraud, reduce errors, and enhance operational efficiency. In the US, AI is streamlining dispute resolution, improving accuracy and reducing turnaround times. The UK’s Faster Payments Service (FPS) is setting global standards for secure, real-time operations with integrated compliance controls. Meanwhile, banks in India, Brazil, and across the EU are exploring AI for real-time risk monitoring and dispute handling. The emergence of Agentic AI adds a predictive, self-learning layer that can autonomously detect, decide, and act transforming RTP ecosystems globally. Agentic AI brings autonomous decision-making, enabling systems to detect, act, and learn driving smarter fraud prevention and faster, context-aware dispute resolutions. To unlock RTP’s full potential, banks must center their dispute strategy around AI backed by regulatory alignment, robust assurance mechanisms, and proactive consumer education to ensure secure, compliant, and future-ready operations dispute management in the real-time era.
AmEx’s integration with Navan to enable business users to instantly create unique virtual cards for travel bookings with built-in spending policies while offering automated reconciliation and real-time expense management
Navan announced a new integration with American Express that enables American Express U.S. Business and Corporate Card Members to instantly create unique virtual Cards for travel booked on the Navan Travel platform via Navan Connect. Navan Connect’s “Bring Your Own Card” functionality enables businesses to enjoy the benefits of the travel and expense solution employees love while keeping the benefits of the company’s existing bank and corporate card partner. To support and foster this integration, Navan is participating in the American Express Sync Commercial Partner Program. Combined with the end-to-end Navan T&E solution, the Navan-American Express Sync integration offers: Improved reconciliation. Speed up month-end close with automated reconciliation, all while earning the rewards of your American Express Card. Proactive spending policies. Create unique virtual Cards with built-in spending policies that make managing travel spend simple for finance teams. Real-time expense management. Companies have full visibility into every virtual Card expense the instant it happens with pending and cleared transactions that automatically appear in the Navan Expense dashboard to enable finance leaders to uncover savings opportunities — while keeping budgets and forecasts up-to-date. With Navan there are even more reasons to love your Card. American Express Card Members can earn the rewards of their eligible American Express Card when they use on-demand virtual Cards for travel payments.
A hacker was able to infiltrate a plugin for an Amazon generative AI assistant after obtaining stolen credentials and making unauthorized changes, including secretly instructing it to delete files
Coders who use artificial intelligence to help them write software are facing a growing problem, and Amazon.com Inc. is the latest company to fall victim. A hacker was recently able to infiltrate a plugin for an Amazon generative AI assistant1 after obtaining stolen credentials and making unauthorized changes, including secretly instructing it to delete files from the computers it was used on. The incident points to a gaping hole in the security practices of AI coding tools that has gone largely unnoticed in the race to capitalize on the technology. The hacker effectively showed how easy it could be to manipulate artificial intelligence tools — through a public repository like Github — with the the right prompt. Amazon ended up shipping a tampered version of the plugin to its users, and any company that used it risked having their files deleted. Fortunately for Amazon, the hacker deliberately kept the risk for end users low in order to highlight the vulnerability, and the company said it “quickly mitigated” the problem. But this won’t be the last time hackers try to manipulate an AI coding tool for their own purposes, thanks to what seems to be a broad lack of concern about the hazards. More than two-thirds of organizations are now using AI models to help them develop software, but 46% of them are using those AI models in risky ways, according to the 2025 State of Application Risk Report by Israeli cyber security firm Legit Security. “Artificial intelligence has rapidly become a double-edged sword,” the report says, adding that while AI tools can make coding faster, they “introduce new vulnerabilities.” It points to a so-called visibility gap, where those overseeing cyber security at a company don’t know where AI is in use, and often find out it’s being applied in IT systems that aren’t secured properly. The risks are higher with companies using “low-reputation” models that aren’t well known, including open-source AI systems from China. Dive into the shadow world of hackers and cyber-espionage. The flaw was discovered by the Swedish startup’s competitor, Replit; Lovable responded on Twitter by saying, “We’re not yet where we want to be in terms of security.” One temporary fix is — believe it or not — for coders to simply tell AI models to prioritize security in the code they generate. Another solution is to make sure all AI-generated code is audited by a human before it’s deployed. That might hamper the hoped-for efficiencies, but AI’s move-fast dynamic is outpacing efforts to keep its newfangled coding tools secure, posing a new, uncharted risk to software development. The vibe coding revolution has promised a future where anyone can build software, but it comes with a host of potential security problems too.
New ‘persona vectors’ from Anthropic helps to identify, monitor and control character traits in LLMs before the models can develop undesirable personalities (e.g., becoming malicious, excessively agreeable, or prone to making things up)
A new study from the Anthropic Fellows Program reveals a technique to identify, monitor and control character traits in large language models (LLMs). The findings show that models can develop undesirable personalities (e.g., becoming malicious, excessively agreeable, or prone to making things up) either in response to user prompts or as an unintended consequence of training. The researchers introduce “persona vectors,” which are directions in a model’s internal activation space that correspond to specific personality traits, providing a toolkit for developers to manage the behavior of their AI assistants better. In a series of experiments with open models, such as Qwen 2.5-7B-Instruct and Llama-3.1-8B-Instruct, the researchers demonstrated several practical applications for persona vectors. A key application for enterprises is using persona vectors to screen data before fine-tuning. The researchers developed a metric called “projection difference,” which measures how much a given training dataset will push the model’s persona toward a particular trait. This metric is highly predictive of how the model’s behavior will shift after training, allowing developers to flag and filter problematic datasets before using them in training. For companies that fine-tune open-source models on proprietary or third-party data (including data generated by other models), persona vectors provide a direct way to monitor and mitigate the risk of inheriting hidden, undesirable traits. The ability to screen data proactively is a powerful tool for developers, enabling the identification of problematic samples that may not be immediately apparent as harmful. The research found that this technique can find issues that other methods miss, noting, “This suggests that the method surfaces problematic samples that may evade LLM-based detection.”
Study says while paying down debt remains a top priority for many households, a significant number are failing to take concrete steps that could ease financial strain and accelerate repayment.
A study from consumer finance platform Happy Money has revealed a striking disconnect in how Americans think about debt versus how they manage it. While paying down debt remains a top priority for many households, a significant number are failing to take concrete steps that could ease financial strain and accelerate repayment. More than one-third of respondents ranked debt repayment among their leading financial goals, yet one in five admitted they had taken no action in the past six months to address it. Only a small fraction had consolidated or refinanced debt strategies that could reduce interest costs and shorten payoff timelines. Meanwhile, with average credit card APRs now exceeding 20%, the cost of carrying a balance is more punishing than ever, and over a third of cardholders continue to roll debt from month to month. The consequences of this inaction reach beyond personal finances. The report highlights a direct link between debt-related concerns and well-being: more than 40% of those worried about credit card payments reported an impact on their mental health, and a third said it disrupted their sleep. Middle-aged consumers appear to be feeling the pinch most acutely, with nearly half in the 35–54 age bracket carrying a balance every month. Happy Money CEO Matt Potere believes the solution lies in making responsible borrowing more accessible and better understood. He argues that creditworthy borrowers often overlook the benefits of fixed-rate personal loans, which can replace high-interest revolving credit with predictable monthly payments. This, he says, can help consumers pay off multiple cards faster, save money on interest, and even improve their credit scores. For the FinTech sector, the findings present a clear opportunity. As traditional lenders and digital challengers alike search for ways to attract and retain customers, offering transparent, affordable, and wellness-focused credit solutions could prove a powerful differentiator.