UK financial regulator partners with Nvidia in AI 'sandbox'
LONDON -Financial firms in Britain will be able to test artificial intelligence tools later this year in a regulatory "sandbox" launched on Monday by the country's financial watchdog, part of a broader government strategy to support innovation and economic growth.
The Financial Conduct Authority (FCA) has partnered with U.S. chipmaker Nvidia to provide access to advanced computing power and bespoke AI software through what it calls a "Supercharged Sandbox."
A sandbox refers to a controlled environment where companies can test new ideas such as products, services or technologies.
The programme is intended to help firms in the early stages of exploring AI, offering access to technical expertise, better datasets and regulatory support, the FCA said. It is open to all financial services companies experimenting with AI.
"This collaboration will help those that want to test AI ideas but who lack the capabilities to do so," Jessica Rusu, the FCA's chief data, information and intelligence officer, said. "We'll help firms harness AI to benefit our markets and consumers, while supporting economic growth."
Finance minister Rachel Reeves has urged Britain's regulators to remove barriers to economic growth, describing it as an "absolute top priority" for the government.
In April, she said she was pleased with how the FCA and the Prudential Regulation Authority, part of the Bank of England, were responding to her call to cut red tape.
Nvidia said the initiative would allow firms to explore AI-powered innovations in a secure environment, using its accelerated computing platform.
"AI is fundamentally reshaping the financial sector," said Jochen Papenbrock, EMEA head of financial technology at Nvidia, citing improvements in data analysis, automation and risk management.
He added that the sandbox will provide firms with a "secure environment to explore AI innovations using Nvidia's full-stack accelerated computing platform, supporting industry-wide growth and efficiency."
The testing is set to begin in October.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Forbes
23 minutes ago
- Forbes
The Dupe Unicorn: The Rise Of Lean, AI-Powered Startups
Wayne Liu is the Chief Growth Officer and President of Perfect Corp America. Venture capital has entered a new era of constraint. According to EY, a $40 billion AI transaction helped prop up venture capital numbers, but without it, funding would have seen a steep 36% drop—and the overall volume of deals continues to shrink. Megadeals like this are now the exception, not the norm. Yet amid this pullback, one sector continues to surge: artificial intelligence. AI-related startups accounted for 33% of all VC activity in 2024 and show no signs of slowing. The unicorn model may be fading, but something smarter is rising in its place. According to Crunchbase, IPO activity has flatlined, with hundreds of companies still in the pipeline; many from the unicorn class of 2021 and 2022 are stuck without an exit in sight. At the same time, traditional venture capital is morphing into something more conservative. As Bloomberg recently reported, firms like Lightspeed are shifting away from the classic VC model, increasingly behaving like private equity players focused on capital efficiency and profitability. In this environment, the billion-dollar unicorn model, the holy grail of startup success for decades, no longer feels viable for many founders. The good news? There seems to be something new taking its place. From the wreckage of the old growth-at-all-costs model, a new breed of startup is emerging: lean, AI-powered and profitable, which is what I refer to as the LAP model. These founders aren't scaling head count or chasing inflated valuations. Instead, they're building lean businesses with small teams, using open-source large language models (LLMs) and AI tools to automate tasks ranging from customer service and content creation to coding and strategic decision-making. In some cases, they're even replacing expensive consultants with AI agents that can perform real-time market analysis or draft legal documents. Recent research shows that AI startups are achieving revenue growth at a pace that significantly outpaces traditional tech companies. Their ability to reach customers faster, automate internal processes and scale with minimal human overhead gives them a notable advantage, especially in a market where efficient monetization has become a survival skill. I call these startups "dupe unicorns." Much like the consumer trend where savvy shoppers seek affordable "dupes" of premium products, today's smartest founders are creating companies that deliver the value of unicorns without the overhead. They may not have massive valuations but offer real utility, faster execution and sustainable business models. In beauty and fashion culture, a "dupe" (short for duplicate) is a lower-cost alternative to a high-end product that offers similar benefits. Crucially, a dupe is not a counterfeit; it doesn't pretend to be the original or infringe on intellectual property. Instead, it's a smart substitute: an independent product that delivers comparable results, often at a fraction of the price. On platforms like TikTok, dupe culture is a celebration of value, discovery and democratized access. Apply that to startups, and the metaphor becomes powerful. A dupe unicorn isn't a knockoff of a legacy tech giant. It's a next-gen business that mimics unicorns' performance, innovation and market traction but with leaner resources, smarter automation and sharper focus. I've seen this shift firsthand through my work advising university accelerators and incubators. Founders today aren't waiting for large teams or fundraising milestones. One founder built an AI-powered video analytics system for sports, helping coaches with limited staff make better decisions. His tool rivals those used by elite organizations at a fraction of the cost. Another, frustrated by traditional market research, launched a generative AI search engine using open-source models, making real-time consumer insights as intuitive as a Google search. A gamer trained a translation engine for cross-border multiplayer sessions, optimized for speed and slang. A marketer developed an AI content generation tool to produce campaign materials, social posts and email copy all from one interface, bypassing costly agencies. It's a new kind of builder mindset: practical, fast-moving and unburdened by the myths of needing massive teams or massive rounds to make an impact. Having helped build my company from seed stage to IPO over nine years, I've lived through the unicorn arc. I know the exhilaration of hypergrowth and its weight. That perspective has sharpened my appreciation for what these new AI-native startups are doing. Here's what I believe founders should embrace now: 1. Design AI-inclusive models from day one. Use AI not just to optimize, but to reimagine your business's foundation. Ask yourself: What can be automated, personalized or accelerated from day one? 2. Run lean with intentionality. It's not about avoiding people; it's about sequencing. Use AI to build early momentum and add team members when human creativity, trust or judgment are truly needed. 3. Speak to the new consumer journey. Gen Z and Gen Alpha founders already get this. The customer journey is no longer a funnel; it's dynamic, nonlinear, shaped by TikTok trends, AI recommendations and immersive technologies like AR glasses. Your go-to-market strategy needs to flow with that reality, not fight it. 4. Prototype, learn, repeat. Founders can now test a new idea, build a landing page, generate content and gather user feedback all within a weekend. Use that velocity to your advantage. 5. Prioritize profitability without compromising vision. Capital efficiency is a competitive advantage. Let profitability guide your build decisions. A good business model should sustain you, not distract you. The unicorn isn't dead but evolving. In today's climate, founders don't need a million-dollar funding and a 30-person team to build a game-changing company. Sometimes, all it takes is a strong idea, a solid AI stack and a sharp understanding of your customer's real need. The dupe unicorn is not a compromise. It's a smarter model for a smarter age, rooted in innovation and built to last. In the age of capital discipline, the LAP model, lean, AI-powered, and profitable, may well become the blueprint for the next generation of breakout startups. Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
Yahoo
25 minutes ago
- Yahoo
Nine million pensioners to receive winter fuel payment this year, Reeves says
Some nine million pensioners in England and Wales will receive the winter fuel payment this winter, the Chancellor has announced. The payment, worth up to £300, will be restored to the vast majority of pensioners who previously received it because anyone with an income of under £35,000 a year will now get the payment automatically. Those with an income above this threshold will also receive the payment, but it will then be reclaimed from them in tax. To be eligible for the winter fuel allowance, a person will need to have reached state pension age by the week starting September 15 this year. Devolved authorities in Scotland and Northern Ireland will each receive a funding uplift so they too can meet the new threshold. Pensioners who do not want to receive the payment will be able to opt out, according to the Treasury. The decision to limit the winter fuel payment to only those who claimed pension credit was one of Labour's first acts in Government, aimed at balancing what was described as a £22 billion 'black hole' in the public finances. This meant the number of pensioners receiving the payment was reduced by around 10 million, from 11.4 million to 1.5 million. But Sir Keir Starmer announced there would be a partial U-turn on the policy in May, after it was thought to have contributed to Labour's drubbing in the local elections. The Treasury claims the new arrangement will cost £1.25 billion in England and Wales, while means-testing winter fuel will save the taxpayer £450 million. Chancellor Rachel Reeves said: 'Targeting winter fuel payments was a tough decision but the right decision because of the inheritance we had been left by the previous government. 'It is also right that we continue to means test this payment so that it is targeted and fair, rather than restoring eligibility to everyone including the wealthiest. 'But we have now acted to expand the eligibility of the winter fuel payment so no pensioner on a lower income will miss out. 'This will mean over three-quarters of pensioners receiving the payment in England and Wales later this winter.' Some two million pensioners who earn more than £35,000 will see their winter fuel payments clawed back via the taxman, the Treasury estimates. Kemi Badenoch, Leader of the Opposition, claimed the Prime Minister had 'scrambled to clear up a mess of his own making'. The Conservative leader added: 'I repeatedly challenged him to reverse his callous decision to withdraw winter fuel payments, and every time Starmer arrogantly dismissed my criticisms. 'This humiliating U-turn will come as scant comfort to the pensioners forced to choose between heating and eating last winter. The Prime Minister should now apologise for his terrible judgment.' Liberal Democrat Leader Sir Ed Davey said: 'Finally the Chancellor has listened to the Liberal Democrats and the tireless campaigners in realising how disastrous this policy was, but the misery it has caused cannot be overstated. 'Countless pensioners were forced to choose between heating and eating all whilst the Government buried its head in the sand for months on end, ignoring those who were really suffering. 'We will now study the detail of this proposal closely to make sure those who need support actually get that support. The pain they went through this winter cannot be for nothing.'
.jpg&w=3840&q=100)
WIRED
37 minutes ago
- WIRED
It's Time to Kill Siri
After almost 10 years, Google Assistant was recently axed in favor of Gemini. Siri gets a bad rap—now is the time for Apple to make a change too. If you've had a passing interest in Apple over the past year, you've likely heard of the company's struggles in the AI race. Apple Intelligence, which arrived slightly late after the launch of the iPhone 16, fell short of expectations, and Apple has yet to deliver the much-improved Siri it promised at WWDC 2024. Siri got a new look and an integration with ChatGPT, but its ability to understand your personal context via emails, messages, notes, and calendar was 'indefinitely' delayed earlier this year as Apple is reportedly facing several challenges. Even if Apple delivered a better Siri, would people use it? Despite arriving first, Siri has long been derided by iPhone owners, often the butt of a joke, as Google Assistant and Alexa rose to the top. But if Apple wants its customers to take the supposed improvements coming to the voice assistant seriously, it should consider taking a page from Google and killing it off for something new. Google has no problem with pulling the plug when things aren't working, or priorities change. In fact, the search giant has a history of killing so many of its services that there's a website dedicated to tracking all the gravestones. One of its most recent terminations? Google Assistant. Photograph: Julian Chokkattu Nearly 10 years since its debut, Google Assistant is in the process of being phased out from every ecosystem it was a part of. Wear OS smartwatches? It's being replaced soon. Android Auto? In the coming months. It's already no longer the default assistant on Android phones. By 2026, it's unlikely we'll see the branding anywhere anymore. So ends the reign of arguably the most effective voice assistant of its time, gone without a care in the world. But Google's decision to kill it, instead of keeping the Google Assistant name, may have been smart. 'It's primarily branding,' says Chris Harrison, who directs the Future Interfaces Group at Carnegie Mellon's Human-Computer Interaction Institute. 'But it underlines a technology reason, which is that the previous generation of these assistants really weren't very much like assistants. Asking for the weather and setting a timer—not very sophisticated. You wouldn't really ask a personal assistant for those mundane tasks.' Gemini is completely different. It can rummage through your emails to find the location of your kid's soccer match, parse through large documents, and when paired with a camera-enabled device, can understand what you're seeing and offer help. Its capabilities are vastly superior to what Google Assistant could do. Apple's goal is to achieve similar results in a more privacy-friendly way—so that when you have Siri connect to ChatGPT, your data is not passed off to OpenAI. 'Apple thought Siri's capabilities would grow, but that didn't really materialize; Siri kind of atrophied out of the gate," Harrison says. 'Now, we're in this new generation of things that are really much more like assistants—they can do reasoning, personalization.' But while Google Assistant and Gemini both have voice interfaces, and at a first glance, they may share a similar look, they're two different applications. 'Simply renaming it Google Assistant 2.0 would not spur people to use it in a fundamentally different way.' It seems that switch to Gemini has been key to moving customer understanding along. However, it's fair to say that Apple and even Amazon's Alexa have had a cultural cachet that Google Assistant never enjoyed. It wasn't unusual to hear Siri or Alexa's name in a movie or TV show; they were much more recognizable than Google's generic-named voice assistant. This may be why Amazon decided to keep the Alexa branding and simply add a '+' icon to denote the new souped-up version of Alexa powered by the latest large language models—and perhaps why Apple is still hanging onto Siri. Photograph: Julian Chokkattu This might have all been OK if Apple actually delivered on its promise and released a functioning, much-improved Siri when it originally said it would. With a massive marketing push to put Apple Intelligence in everyone's mind (maybe a regretful move), it would have been a great opportunity to wow users with a much-improved Siri. Months later, customers are left wondering why Siri—new look and all—still lags behind. But the broader problem affecting all large language models isn't just the branding, but the user interface. Harrison compares it to the days of command-line computing and the shift to the graphical user interface (GUI) in the '80s and '90s. It wasn't the graphics that made the latter more popular, but the discoverability and explorable interface. In the command-line era, you had to remember how to do anything. With GUI, you could put anyone in front of a computer, and they'd be able to figure out how to navigate the operating system. If you put someone in front of ChatGPT or Gemini, say it's an incredible tool, and tell them to ask it anything, they'll just stare blankly at the blinking prompt. 'It's like we've gone back 30 years in interface design. They have no idea what to do or say." Harrison says he did this exact experiment with his parents: They asked what the weather was tomorrow, and the AI responded that it didn't have that information. 'We've regressed in discoverability," he says. 'A regular person, not the tech people, if all they've been doing is setting timers with Siri for the past 10 years, and now they have to think about it in a fundamentally different way—that's an extremely hard problem. Some sort of renaming of the application is going to be important." Saying goodbye to Siri would be a big move for Apple—after all, it has spent more that a decade investing in it. But most people today still use it for playing music, checking weather, and setting timers, and aren't even pushing the boundaries of its current, relatively limited, capabilities. It's hard to see that changing anytime soon, even if Siri's feature-packed next generation arrives as promised. 'For 99 percent of the planet, this kind of AI revolution has totally gone over their head," Harrison says. Like the 10-year transition from command line to graphical user interfaces, rethinking the way we use these personal voice assistants will take time and education, but maybe a new name will help Apple with the transition.
