Klarna's AI replaced 700 workers — Now the fintech CEO wants humans back after $40B fall
Klarna Group Plc's co-founder and CEO, Sebastian Siemiatkowski, has admitted the fintech giant's aggressive use of artificial intelligence in customer service has backfired. 'As cost unfortunately seems to have been a too predominant evaluation factor when organizing this, what you end up having is lower quality,' he said at Klarna's Stockholm headquarters.
The company had halted hiring for over a year to focus on building AI capabilities, part of a broader cost-cutting effort. However, Siemiatkowski now says the shift needs recalibration. 'Really investing in the quality of the human support is the way of the future for us.'
In a strategic pivot, Klarna is launching a fresh recruitment drive for customer support roles — a 'rare' move, according to a report in Bloomberg. The firm is piloting a new model where remote workers, such as students or people in rural areas, can log in and provide service on-demand, 'in an Uber type of setup.' Currently, two agents are part of the trial.
'We also know there are tons of Klarna users that are very passionate about our company and would enjoy working for us,' Siemiatkowski said. He emphasised that from both a 'brand perspective' and 'company perspective,' it is critical to 'always' give customers the option to speak to a human.
While scaling back its all-in AI push for customer service, Klarna remains committed to integrating artificial intelligence across its operations. The firm is rebuilding its technology stack with AI at the core to drive efficiency, and is working on a digital financial assistant to help customers secure better interest rates and insurance deals, the news outlet reported.
Siemiatkowski, 43, said Klarna's relationship with OpenAI remains strong. 'We wanted to be [OpenAI's] favorite guinea pig,' he recalled, referencing Klarna's early collaboration with the AI leader in 2023.
Klarna's renewed hiring comes after a period of turbulence. The company's valuation plunged from a peak of $45.6 billion in 2021 to $6.7 billion during a 2022 funding round. It has since rebounded, with plans to raise $1 billion via IPO at a valuation of over $15 billion — though those plans were recently paused amid market volatility, according to the news report.
The company's 2024 announcement that AI was handling the workload of 700 human agents shook the call center industry, sending shares of France-based Teleperformance SE down sharply.
Despite the return to hiring, Klarna's headcount is still expected to shrink through attrition. 'In a year's time,' Siemiatkowski said, 'we'll probably be down to about 2,500 people from 3,000.' He added the pace of downsizing could increase as AI technology improves.
'I feel a bit like Elon Musk,' the Klarna CEO quipped, 'always wanting to say it's going to happen tomorrow, when it's going to take a little bit longer. I think it's very likely within 12 months.'
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Economic Times
12 minutes ago
- Economic Times
Is ChatGPT making us dumb? MIT brain scans reveal alarming truth about AI's impact on the human mind
MIT researchers have discovered that using ChatGPT for essay writing reduces brain engagement and learning over time. Through EEG brain scans of 54 students, those who relied on AI performed worse than others across neural and linguistic metrics. The study raises concerns that AI tools may hinder critical thinking and promote passive acceptance of algorithm-driven content. A new MIT study warns that regular use of ChatGPT could impair memory, brain activity, and critical thinking. Students relying on the AI tool showed significantly lower cognitive engagement than peers using Google or no tools at all. (Representational image: iStock) Tired of too many ads? Remove Ads Brain vs Bot: How the Study Was Done Google Wasn't Great, But Still Better Than ChatGPT Tired of too many ads? Remove Ads A Shortcut with a Hidden Toll What This Means for the AI Generation It's quick, it's clever, and it answers almost everything—no wonder millions around the world rely on ChatGPT . But could this digital genie be dulling our minds with every wish we make? According to a startling new study by scientists at MIT's Media Lab, the answer may be yes. Researchers have now found that excessive use of AI tools like ChatGPT could be quietly eroding your memory, critical thinking, and even your brain on arXiv, the study titled 'The Cognitive Cost of Using LLMs' explores how language models—especially ChatGPT—affect the brain's ability to think, learn, and retain examine what they call the 'cognitive cost' of using large language models (LLMs), MIT researchers tracked 54 students over a four-month period using electroencephalography (EEG) devices to monitor brain activity. The participants were divided into three groups: one used ChatGPT, another relied on Google, and the last used no external help at all—dubbed the 'Brain-only' the AI-powered group initially showed faster results, the long-term findings were more sobering. Students who depended on ChatGPT for essay writing exhibited poorer memory retention, reduced brain engagement, and lower scoring compared to their peers. As the researchers noted, 'The LLM group's participants performed worse than their counterparts in the Brain-only group at all levels: neural, linguistic, and scoring.'Interestingly, students who used Google showed moderate brain activity and generated more thoughtful content than those who leaned on ChatGPT. Meanwhile, those in the Brain-only group had the highest levels of cognitive engagement, producing original ideas and deeper insights. In fact, even when ChatGPT users later attempted to write without assistance, their brain activity remained subdued—unlike the other groups who showed increased engagement while adapting to new suggests that habitual ChatGPT usage might not just affect how we think, but whether we think at study also points to how this over-reliance on AI encourages mental passivity. While ChatGPT users reported reduced friction in accessing information, this convenience came at a cost. As the researchers explained, 'This convenience came at a cognitive cost, diminishing users' inclination to critically evaluate the LLM's output or 'opinions'.'The team also raised red flags about algorithmic bias : what appears as top-ranked content from an AI is often a result of shareholder-driven training data, not necessarily truth or value. This creates a more sophisticated version of the 'echo chamber,' where your thoughts are subtly shaped—not by your own reasoning, but by an AI's probabilistic AI tools become more embedded in our everyday tasks—from writing emails to crafting essays—this study is a wake-up call for students, educators, and professionals. While tools like ChatGPT are powerful assistants, they should not become cognitive researchers caution that as language models continue to evolve, users must remain alert to their potential mental side effects. In a world where convenience is king, critical thinking might just be the first casualty.
Hans India
33 minutes ago
- Hans India
OpenAI Scrubs Jony Ive Deal from Web Amid Trademark Battle with Rival AI Firm
In a sudden turn of events, OpenAI has removed all public references to its much-discussed $6.5 billion acquisition of Jony Ive's AI hardware startup, IO Products. The move follows a legal challenge over trademark infringement by a competing firm with a similar name, stirring speculation and concern across the tech community. Over the weekend, social media users noticed the disappearance of promotional content and the official website linked to the IO Products deal. Both OpenAI and representatives for Jony Ive have since confirmed that the takedown was prompted by an ongoing trademark dispute with IYO Inc., a company also operating in the AI device space. According to Bloomberg, the case is now under judicial review. Responding to the controversy, a spokesperson for Ive called the legal complaint 'an utterly baseless complaint and we'll fight it vigorously.' The IO Products deal, officially unveiled in May, marked a pivotal shift for OpenAI—best known for its generative AI software. The partnership with Ive, the legendary former Apple designer behind iconic devices like the iPhone and iMac, aimed to bring OpenAI's first AI hardware device to market within a year. At the time of the announcement, OpenAI CEO Sam Altman had predicted that the company could sell up to 100 million units of the new hardware. He even suggested the acquisition could potentially boost OpenAI's valuation by $1 trillion, stating, 'This is the biggest thing we've ever done as a company.' In a now-deleted blog post, Altman and Ive had jointly written, 'It became clear that our ambitions to develop, engineer and manufacture a new family of products demanded an entirely new company.' OpenAI's financial commitment to the deal included $5 billion in equity, in addition to its existing 23% stake in IO from an earlier collaboration. IO Products was incorporated in Delaware in September 2023 and later registered in California in April 2025, according to public filings. Heading the new hardware division is OpenAI executive Peter Welinder, known for pioneering work in robotics and innovative product development. The team includes experts across design, hardware, software, and manufacturing, collaborating closely with OpenAI's San Francisco-based engineering and research teams. Altman and Ive's collaboration dates back to early 2023, though the public got its first hint of an 'AI-first device' in February 2025. While specific details remain under wraps, industry watchers speculate the product could rival the Humane AI Pin or Rabbit R1, or even explore futuristic formats like smart glasses, in-car systems, or AI-integrated robots. Reflecting on the collaboration, Ive shared, 'I have a growing sense that everything I have learned over the last 30 years has led me to this moment. I am so grateful for the opportunity to be part of such an important collaboration.' Altman added his admiration for Ive's approach, saying, 'AI is an incredible technology, but great tools require work at the intersection of technology, design, and understanding people and the world. No one can do this like Jony and his team; the amount of care they put into every aspect of the process is extraordinary.' As legal proceedings continue, OpenAI's ambitious foray into hardware hangs in the balance, with stakeholders closely watching how the trademark challenge might affect the future of this high-stakes collaboration.
Time of India
an hour ago
- Time of India
Trouble between AI's power couple: What's brewing between Microsoft and OpenAI?
The biggest partnership in the artificial intelligence (AI) world, between Microsoft and OpenAI , is showing signs of cracks. Microsoft invested $1 billion in the ChatGPT maker in 2019, ahead of the generative AI boom . The companies then rode the wave to the top, with the software maker injecting billions more in the company in the following years. But then disagreements arose over controlling AI technology and computing resources, intellectual property rights, as well as OpenAI's organisational transition plans and competitive tensions. These issues have made the once-close partnership shaky, and may fundamentally change it. What's happened? Live Events OpenAI needs Microsoft's approval to complete its transition into a public-benefit corporation. But sources told Reuters the companies have not been able to agree on details even after months of negotiations. Discover the stories of your interest Blockchain 5 Stories Cyber-safety 7 Stories Fintech 9 Stories E-comm 9 Stories ML 8 Stories Edtech 6 Stories Tensions have risen between the two sides after reports emerged that OpenAI is considering a "nuclear option", accusing Microsoft of anticompetitive behaviour. The companies are discussing revising the terms of Microsoft's investment, including the future equity stake it will hold in OpenAI, the report said, adding that the ChatGPT owner wants to modify existing clauses that give Microsoft exclusive rights to host OpenAI models in its cloud. Reuters reported that Microsoft was even ready to walk away from its high-stakes negotiations with OpenAI over the future of its alliance. Windsurf issue The key issue in the dispute is now Microsoft's access to OpenAI's intellectual property. OpenAI even wanted to exclude Microsoft's access to AI coding startup Windsurf, which it had acquired, due to competing products. OpenAI acquired Windsurf, an AI-assisted coding tool formerly known as Codeium, for about $3 billion, marking the company's largest acquisition to date. Microsoft-owned Github offers a competing AI tool for programmers. Investors have also poured money into a new crop of startups offering similar tools, including Anysphere, the startup behind Cursor. Analysts have said that the partnership between the legacy giant and the AI startup was always unstable, with Microsoft testing alternatives and preparing for a way forward without OpenAI.
