Latest news with #AI-cloned
%2520(1).png&w=3840&q=100)
WIRED
06-05-2025
- Business
- WIRED
Turning The Tide: How Mastercard is Pushing Back on Fraudsters and Reducing Transaction Confusion
Criminal networks are deploying increasingly sophisticated methods, costing consumers billions. With AI helping to protect its 159 billion transactions a year, Mastercard is turning the tide and safeguarding trust Whether it's fake delivery texts, fraudulent online storefronts, or criminals using AI-cloned voices to fool your bank's phone security systems, fraudsters are getting more sophisticated—and bolder. More than half of recently polled consumers reported spotting a fraud attempt at least once a week. It's no wonder UK consumers lost an estimated £1.17 billion to fraud in 2023. Few understand the scale of the challenge—or the path towards solutions—as well as Mastercard. Since the earliest days of credit card fraud, the global technology company, which today processes 159 billion payments a year, has been going toe-to-toe with criminal gangs, examining trends and identifying innovative new ways to outsmart fraudsters. Now, with criminal networks increasingly turning to AI and other weapons to deceive their victims, Mastercard is deploying its own technology to fight back. 'Much of our innovation and our readiness to confront fraud is borne out of our investment in technology and data,' explains Johan Gerber, Executive Vice President and Head of Mastercard's Security Solutions business. 'This is what drives our understanding of the patterns and trends that help us become more resilient.' Protecting the spectrum of payments It starts from the moment you open an account. Mastercard's AI tools can spot fraudsters attempting to use synthetic or fake identities. It can determine whether a customer is genuine by the way they hold their phone or swipe up or down. When you attempt a new transaction, the company's Decision Intelligence technology—trained on hundreds of billions of transactions and trillions of data points, including account, purchase, merchant, and device information—analyzes in just 50 milliseconds whether the behavior is suspicious and therefore likely to be fraudulent. Mastercard says its recent enhancements to the technology is working, detecting more than 40% more fraud versus Q1 last year. And when you do make a purchase, Mastercard's tech provides detailed and transparent breakdowns of each purchase in your banking app, so that you can quickly review transactions to check for potential fraud. Smarter, More Secure, and More Personal As ecommerce transactions skyrocket, some cardholders are challenging genuine transactions. This trend, known as First Party Fraud, is already estimated to cost merchants around $50bn (£39bn) per year. To help combat this growing fraud type, Mastercard's First-Party Trust program, already live in the US, leverages advanced identity technology and AI to enable businesses to fight instances where genuine transactions are mistakenly or intentionally challenged by cardholders. As a result, Mastercard has observed a 15-20 percent chargeback deflection rate in initial studies. 'As ecommerce booms, we need to focus not only on reducing fraud, but on structuring the right incentives to encourage proactive investment in security. This needs to be sustainable for all parties,' says Gerber. 'It is part of our vision for keeping payments smarter and more personal, as well as more secure." In some cases, cyber criminals have been able to penetrate banks' own security systems, disabling fraud detection methods and withdrawal limits. In 2018, for example, cyber criminals allegedly used the method to steal £11.5m from ATMs in 28 countries in just 2 hours. As criminals leverage AI more and more, Mastercard has constantly evolved its tech to prevent such widespread attacks by monitoring transactions at the network level, flagging suspicious behavior with banks as they happen. The company claims that doing so has prevented almost $50 billion in thefts in the last two years alone. But a lot of the time, scams are personal: a phone call, text message, or a DM seemingly from a relative, or loved-one. Romance and 'fake relative' scams are just two examples of frauds designed to trick you into sending money willingly. In 2023, so-called Authorised Push Payment (APP) frauds conned UK consumers out of £459 million. To combat this, Mastercard's world-first Consumer Fraud Risk technology analyzes risk factors around a transaction and alerts both the sending and receiving banks. Since 2023, the solution has helped 14 UK banks identify and stop scam payments before they left the victim's account. In a bid to combat other types of fraud, Mastercard has also expanded its generative AI capabilities to be able to determine the full 16-digits of compromised numbers hidden by fraudsters in illegal websites, allowing banks to block them quickly before fraud takes place. There's some evidence that the industry's anti-fraud efforts are working: the UK government's own Payment Systems Regulator reported that the value of APP scams has fallen by over 12 percent since 2023. 'With every step forward in technology, we're not just defending against threats—we're shaping a digital future where innovation and security walk hand in hand,' says Gerber. 'This way we are ensuring trust remains the cornerstone of every interaction.' Discover more at
Newsweek
28-04-2025
- Health
- Newsweek
First Neuralink Patient With ALS Can Now Use AI to Have Conversations
Based on facts, either observed and verified firsthand by the reporter, or reported and verified from knowledgeable sources. Newsweek AI is in beta. Translations may contain inaccuracies—please refer to the original content. The first patient with ALS to receive a Neuralink brain transplant has demonstrated his new ability to communicate with only his mind. Bradford Smith, who is the third human to receive the transplant in history, said on X (formerly Twitter) that he was able to "type with my brain" after being rendered largely immobile and nonverbal by the motor neuron disease. Newsweek reached out to Smith via social media for comment. Why It Matters Amyotrophic lateral sclerosis, also known as ALS, is a motor neuron disease that causes a loss of muscle control and speech, often confining patients to a wheelchair. The development of Elon Musk-owned Neuralink, which began human testing in 2024, is a significant breakthrough in treating the disease, as it appears capable of restoring communication for those who are unable to speak. What To Know In a video uploaded to his X account on Monday, Smith said he was the "third person in the world to receive the Neuralink brain implant." "I'm also the first person with ALS and the first nonverbal, which means that I rely on it for all communication," he said. I am the 3rd person in the world to receive the @Neuralink brain implant. 1st with ALS. 1st Nonverbal. I am typing this with my brain. It is my primary communication. Ask me anything! I will answer at least all verified users! Thank you @elonmusk! — Bradford G Smith (Brad) (@ALScyborg) April 27, 2025 The Neuralink device, which Smith described as about the size of "five U.S. quarters in a stack," is implanted in the skull and connected to the brain's motor cortex, which controls the body's movement, through flexible threads. These threads detect neural activity and transmit the signals wirelessly to external computing systems, where machine learning algorithms interpret them into actionable commands. Neuralink commenced human trials in January 2024 after receiving clearance to do so from the Food and Drug Administration in May 2023. In the video, which Smith narrated using an AI-cloned version of his voice, he said: "I have ALS, a really weird disease that kills the motor neurons that control my muscles but not affecting my mind. My experience has been pretty interesting, starting with a shoulder injury that would not heal and ending up with my current status." Smith, who said he could move only his eyes and required a ventilator to breathe, added, "Neuralink has given me freedom, hope and faster communication." File photo of a spinal implant chip from the Wings for Life Foundation at the University of Texas. This is not the Neuralink chip that Bradford Smith had implanted. File photo of a spinal implant chip from the Wings for Life Foundation at the University of Texas. This is not the Neuralink chip that Bradford Smith had implanted. AP Photo What People Are Saying Brad Smith wrote on X on Monday: "I am the 3rd person in the world to receive the Neuralink brain implant. 1st with ALS. 1st Nonverbal. I am typing this with my brain. It is my primary communication. Ask me anything!" Neuralink said in a news release issued on February 5: "In just the past couple of weeks, the Link has enabled Brad to watch his son win a regional robotics competition, chat with neighbors at the park, and give a talk with a live Q&A session to the youth at his local church. "We're currently working with Brad to design a communication system from the ground up by combining language models and novel neural decoding strategies. Our goal is to allow him to engage in conversations at a more natural speaking pace." What Happens Next Neuralink plans to continue its human clinical trials, writing in February, "We have active trials in the United States and Canada and are connecting with people interested in potential future clinical investigations in the United Kingdom."
Yahoo
21-03-2025
- Entertainment
- Yahoo
Some voice actors embrace AI's potential. Others worry it's coming for their jobs
Like many in his field, voiceover actor Jesse Adam has been watching the rise of AI with a wary eye. "You start hearing better voiceover produced by AI and you're like, OK, this could be some serious competition," says the Saskatchewan native, who over the course of his decade-long career has lent his voice to dozens of projects, including Marvel video games and Starbucks commercials. With AI capabilities advancing at a seemingly exponential rate, he realizes he faces an uphill battle. 'I got to the point where I was like, there's no way we're going to be able to fight this. This isn't going away." So rather than resist the tide, he's decided to lean into it. Adam has long worked with a platform called which connects voice actors with businesses for commercial and corporate work. Last summer, the company, based in London, Ont., launched a new AI Studio initiative, which allows clients to buy clones of real actors' voices for text-to-speech projects. The company says it allows its talent to opt out at any time, and Adam was fully on board with that. He hopes for a future where his voice clone generates passive income and gives him the freedom to pursue passion projects. 'The industry is changing and I need to adapt or I'm going to get left behind," he says. Adam spent roughly 20 hours recording a range of emotional performances for the service — everything from 'happy and upbeat' to 'sad and down.' Since then, he's been paid for about 100 jobs, mostly for training videos. Each time his AI-cloned voice is used, he gets an email notification. 'It's not paying a ton of money, but it's also money that I wouldn't have had otherwise,' says Adam, noting the service pays him about 10 cents per word and he usually makes more working with clients directly. He hopes to scale up as demand for the service grows. "I think for some talent, the amount of work they'll get through their AI clones will be significant and will pay the bills and be lucrative." About 30 voice actors have opted into AI Studio, according to the company's chief technology officer Dheeraj Jalali. He says the platform expanded into AI text-to-speech services after seeing how quickly the technology was evolving. 'We thought, hey, let's use this technology as an enabler instead of being fearful of it,' he says. Jalali explains actors who consent to having their voices used can set their own per-word rates, with the company taking a percentage of the earnings. They can specify words or phrases they don't want their voice to generate, and talent can view the scripts their AI voice was used for after the fact. He says the service is primarily used for training programs, ebooks and website user interfaces. Since AI can't take artistic direction, Jalali notes voice actors are still being hired traditionally for commercials, films and TV shows. 'When it comes to mass media, voiceover actors are still there. I don't see any near future where they're getting replaced,' he says. But that's precisely what has many actors and labour groups sounding the alarm. The entertainment industry has weathered a rough couple of years, with Hollywood strikes and fewer Canadian commissions slowing things down. The animation sector has been hit especially hard. According to the Canadian Media Producers Association, animation production in Canada plummeted by 55 per cent in 2023-24, following a pandemic-era surge that briefly reignited investments. In the midst of this downturn, voice actors are on alert when it comes to AI. As the technology becomes more skilled at mimicking human voices, many fear being pushed aside and watching their craft lose value. Toronto's Gabbi Kosmidis has spent years breathing life into animated characters, and she worries that an illicit, synthetic version of her voice could replace her entirely. 'I know a lot of voice actors who are struggling to find work right now. The industry is slow for a plethora of reasons, and AI has been this scary, looming thing,' says Kosmidis, who voices the lead in 'Night of the Zoopocalypse,' an animated Canadian horror comedy that did not use AI, and is currently in theatres. She fears a near future where companies 'could take your voice and use it forever and ever and ever.' 'It's pretty scary in terms of voiceover work. It may make your job obsolete.' In light of these fears, labour unions in Canada and the U.S. have been working to ensure members have protections against AI in contracts with producers. ACTRA's newly ratified Independent Production Agreement has language giving actors the right to fair compensation and full consent if their voice or likeness are used to create a synthetic performer. Still, 'a collective bargaining agreement is not enough alone to address all concerns on such a rapidly evolving technology,' ACTRA national president Eleanor Noble said in a statement. She added the organization is urging the Liberal government to create 'strong legislation which will protect performers' likenesses and jobs from AI misuse, along with protecting all Canadian workers.' As the government drafts its proposed Artificial Intelligence and Data Act, ACTRA is lobbying to ensure any future legislation would grant performers the right to fair compensation and control over all AI uses of their voice and likeness. AI text-to-speech services signal a 'race to the bottom' where voiceover artists will get paid less and less for their work, says Kunal Sen, creative director of Vancouver-based animation studio Good Bad Habits, which works on a variety of projects, including films and commercials. 'I definitely see a devaluing of work in this sphere. I would never not hire an actual person to do voiceover work,' he says. 'If a brand needs someone to say something in a video and it doesn't have to be Morgan Freeman or some celebrity, now they can just be like, 'OK, let's just cut the budget from this and put it somewhere else.' It's never been easier to do that.' Jalali acknowledges that AI may drive down voiceover rates but argues that it also allows actors to scale their work. 'Voiceover actors have a way bigger opportunity now,' he says. He envisions a near future where actors license their voices to companies, who will 'have the potential to do multiple projects with your voice in different use cases." He says performers can now tell clients: "Hey, here's my voice. Now go scale with this voice and I'll benefit from it, too, because my AI voice can say a million words a day. I can only say maybe a thousand." "So it just scales up in ways that were not possible beforehand.' Kosmidis remains apprehensive about AI, believing it compromises the quality of performances in the long run. 'I personally would never want to listen to a robot, even if that robot sounds very human. Knowing that it's not really human, to me feels weird and disconnected,' she says. Kosmidis questions whether AI can truly capture emotional nuance or deliver humour effectively. However, she concedes it's likely only a matter of time before it masters the art of crying. "Bursting into tears. I mean, who can't do that? In this economy?' This report by The Canadian Press was first published March 21, 2025. Alex Nino Gheciu, The Canadian Press
