 |
|
|
|||||
Daniel Kokotajlo predicted the end of the world would happen in April 2027. In AI 2027 a document outlining the impending impacts of AI, published in April 2025 the former OpenAI employee and several peers announced that by April 2027, unchecked AI development would lead to superintelligence and consequently destroy humanity. The authors, however are going back on their predictions. Now, Kokotajlo forecasts superintelligence will land in 2034, but he doesnt know if and when AI will destroy humanity. In AI 2027, Kokotajlo argued that superintelligence will emerge through fully autonomous coding, enabling AI systems to drive their own development. The release of ChatGPT in 2022 accelerated predictions around artificial general intelligence, with some forecasting its arrival within years rather than decades. These predictions accrued widespread attention. Notably, JD Vance, U.S. vice president, reportedly read AI 2027 and later urged Pope Leo XIV who underscored AI as a main challenge facing humanity to provide international leadership to avoid outcomes listed in the document. On the other hand, people like Gary Marcus, emeritus professor of neuroscience at New York University, disregarded AI 2027 as a work of fiction, even calling various predictions pure science fiction mumbo jumbo. As researchers and the public alike begin to reckon with how jagged AI performance is, AGI timelines are starting to stretch again, according to Malcolm Murray, an AI risk management expert and one of the authors of the International AI Safety Report. For a scenario like AI 2027 to happen, [AI] would need a lot of more practical skills that are useful in real-world complexities, Murray said. Still, developing AI models that can train themselves remains a steady goal for leading AI companies. Sam Altman, OpenAI CEO, set internal goals for a true automated AI researcher by March of 2028. However, hes not entirely confident in the companys capabilities to develop superintelligence. We may totally fail at this goal, he admitted on X, but given the extraordinary potential impacts we think it is in the public interest to be transparent about this. And so, superintelligence may still be possible, but when it arrives and what it will be capable of remains far murkier than AI 2027 once suggested. Leila Sheridan This article originally appeared on Fast Company‘s sister publication, Inc. Inc. is the voice of the American entrepreneur. We inspire, inform, and document the most fascinating people in business: the risk-takers, the innovators, and the ultra-driven go-getters that represent the most dynamic force in the American economy.
Category:
E-Commerce
For most of modern finance, one number has quietly dictated who gets ahead and who gets left out: the credit score. It was a breakthrough when it arrived in the 1950s, becoming an elegant shortcut for a complex decision. But shortcuts age. And in a world driven by data, digital behavior, and real-time signals, the score is increasingly misaligned with how people actually live and manage money. Were now at a turning point. A foundational system, long considered untouchable, is finally being reconstructed by using AIspecifically, advanced machine learning models built for risk predictionto extract more intelligence from existing data. These are rigorously tested, well-governed systems that help lenders see risk with greater nuance and clarity. And the results are reshaping core economics for lenders. THE CREDIT SCORE WASNT BUILT FOR MODERN CONSUMERS Legacy credit scores rely on a narrow slice of information updated at a pace that reflects the black-and-white television era. A single late payment can overshadow years of financial discipline. Data updates lag behind real behavior. And lenders are forced to make million-dollar decisions using a tool that cant see volatility, nuance, or context. A single, generic credit score is a compromise by design. National credit scores are designed to work reasonably well across thousands of institutions, but not optimally for any specific one. That becomes clear when you compare regional differences. A lender in an agricultural region may see very different income seasonality and cash-flow patterns than a lender in a major metro areadifferences that a universal score was never designed to capture. Financial institutions need models built around their actual membership that can adjust to different financial histories and behaviors. That rigidity has created the gap were now seeing across the economy. Consumers feel squeezed, lenders feel exposed, and businesses struggle to grow in a risk environment that looks nothing like the one their scoring tools were built for. Modern machine-learning models give lenders something the score never coulda panoramic view instead of a narrow window. HOW AI CHANGES THE GAME The data in credit files has long been there. Whats changed is the modelingmodern machine learning systems that can finally make full use of those signals. These models can evaluate thousands of factors inside bureau files, not just the static inputs, but the patterns behind them: How payment behavior changes over time Which fluctuations are warning signs versus temporary noise How multiple variables interact in ways a traditional score cant measure This lets lenders differentiate between someone who is truly risky and someone who is momentarily out of rhythm. The impact is profound: more approvals without more losses, stronger compliance without more overhead, and decisions that align with how people actually manage their finances today. For leadership teams, this also means making intentional choices about who to serve and how to allocate capital. Tailored models let institutions focus their resources on the customers they actually want to reach, rather than relying on a one-size-fits-all score. AI FIXES SOMETHING WE DONT TALK ABOUT ENOUGH There’s widespread concern about AI bias, and rightly so. When algorithms aren’t trained on a representative set of data or arent monitored after deployment, this can create biased results. In lending, these models arent deployed on faith; theyre validated, back-tested, and monitored over time, with clear documentation of the factors driving each decision. Modern explainability techniques, now well-established in credit risk, can give regulators and consumers a clearer view into how and why decisions are made. Business leaders should also consider that there is bias embedded in manual underwriting. Human decisionsespecially in high-volume, time-pressured environmentsvary from reviewer to reviewer, case to case, hour to hour. Machine learning models that use representative data, are regularly monitored, and make explainable, transparent decisions, giving humans a dependable baseline. This allows them to focus on exceptions, tough cases, and strategy. THE NEW ADVANTAGE FOR BUSINESS LEADERS The next era of lending will be defined by companies that operationalize AI with discipline, building in strong governance, clear guardrails, and transparency. Those who do will see higher approval rates, lower losses, faster decisions with fewer manual bottlenecks, and fairer outcomes that reflect real behavior, not outdated shortcuts. For the first time in 70 years, were able to bring real, impactful change to one of the most influential drivers in the economy. THE FUTURE ISNT A SCORE, ITS UNDERSTANDING If the last century of lending was defined by a single, blunt number, the next century will be defined by intelligence. By the ability to interpret risk with nuance, adapt to fast-moving economic signals, and extend opportunity to people who have long been underestimated by the system. AI wont make lending flawless. But it gives us the clearest path weve ever had toward a credit ecosystem that is more accurate, more resilient, and far fairer than the one we inherited. And for leaders focused on growth, innovation, and long-term competitiveness, that shift is transformational. Sean Kamkar is CTO of Zest AI.
Category:
E-Commerce
Perusing the grocery aisle in the Westside Market on 23rd Street in Manhattan, you might not even notice the screens. They look just like paper price labels and, alongside a bar code, use a handwriting-style font weve come to associate with a certain merchant folksiness. Theyre not particularly bright or showy. The only clues that theyre not ordinary sticky shelf labels are a barely distinguishable light bulb and, on some, a small QR code. These are electronic shelf labels, chip-enabled screens that some stores are now using to display product prices. Unlike their paper predecessors, the prices arent printed in ink but rendered in pixels, and they can change instantaneously, at any time. The labels also come with additional features. An LED light can switch on to flag something, perhaps a product that needs restocking, explains Vusion, the company that made the labels Westside Market is now using. The QR codes are designed to help customers find more information about a product, or integrate with a personalized shopping list someone might have. Of course, these labels arent just labels, but end-points of a much larger effort to digitize every way we now interface with products. You have a network in the store. You send the information that you want to transmit to the labels, and there you go, says Finn Wikander, the chief product officer at Pricer, another company thats manufacturing ESLs with the hope of making them a fixture of 21st century shopping. Unsurprisingly, electronic shelf labels have become a flashpoint for consumer anxiety. The companies selling the devices, and the stores buying them, say the technology isnt about screwing people over but about making their businesses easier to run. Automating price changes eliminates hours spent replacing labels. It also makes it simpler to respond to new tariffs or account for rising inflation. But in a world spooked by dynamic pricing, electronic shelf labels can look to some like a goblin of digitizationa symptom of late-stage Silicon Valley campaigns to streamline and optimize seemingly all elements of commerce. Even members of Congress have raised suspicions about the technology, arguing that it enables price gouging and discrimination, particularly as it becomes more common in the United States. “Historically, when we thought about brick and mortar stores, prices were relatively stable,” Vicki Morwitz, a Columbia Business School professor who focuses on marketing and consumer behavior, tells Fast Company. “These electronic shelf tags break that assumption which makes pricing feel less stable. Even if average prices aren’t necessarily going up, that shelf instability can become a psychological flash point.” Screenified everything A handful of companies sell this technology as part of broader enterprise software packages. Theres Pricer, a Swedish firm, and Vusion, headquartered in France. Solum operates out of South Korea, and Opticon, known for barcode scanners, is also in the mix. Electronic shelf labels can also be bought, ahem, off the shelf and integrated into a stores Bluetooth networkno enterprise startup required. The pitch for these devices is exactly what unsettles so many shoppers: Electronic shelf labels make it much easier for stores to change prices dynamically and more frequently. The companies that manufacture and deploy these tools say there are legitimate reasons to do so. For example, a store might raise prices if suppliers increase costs, or cut them quickly when a product is nearing its expiration date. ESLs also allow chains to keep prices consistent across locations and respond more quickly to competitors (especially valuable at a time when shoppers are already carrying smartphones to compare prices between stores). Most consumers today are used to either doing their own scanning or use ChatGPT or Gemini to find the best offer or use price comparison sites, says Pricer’s Wikander. Then theres labor. Employees might spend hours replacing labels for a price surge or sale. The idea is to liberate people from very tedious tasks in a store. Changing prices could be one. Launching promotions could be one, argues Loc Oumier, a marketing executive with Vusion. There are also regulatory considerations: France, for instance, passed a law mandating that prices at checkout match advertised prices on aisles, which pushed stores in that country to adopt the technology, says Wikander. They are now rolling out more broadly in the United States, especially at large chains. Vusion says its labels are in use at Fresh Market, Mattress Firm, and Leons in Canada. Walmart, which declined to comment for this story, announced in 2024 that it would begin installing electronic shelf labels, with plans to bring Vusions technology to more than 2,000 stores by the end of 2026. Tests or deployments have appeared in Whole Foods, Schnucks, and even smaller retailers like Westside Market. The reception can be frosty. While there are some scenarios, like from Uber rides and airline tickets, where consumers have come to accept rapidly changing costs, the practice often feels jarring. That tension was evident in 2024, when Wendys faced backlash after announcing plans to install digital menu boards and later promised it wouldn’t introduce surge pricing for burgers. Shoppers also worry about price gouging, where retailers spike prices during emergencies. Exploiting consumers when they have no real alternatives or limited alternatives, says Columbia’s Morwitz. The problem is consumers may feel exploited long before an economist would say they are. There is also the understandable anxiety that the technology is designed to cut jobs. Some workers, as reported in The Nation, say the labels do not simplify their work but replace one kind of labor with another form of algorithmic babysitting. Unlike paper tags, screens can break, and computer programs fall victim to bugs and internet outages. Employees at one chain store operated by Kroger, which has also deployed the tech, have apparently complained that the labels heat up stores. (Kroger did not respond to Fast Company‘s request for comment.) Concerns reach D.C. Lawmakers have taken notice. Democratic Senators Elizabeth Warren of Massachusetts and Bob Casey of Pennsylvania wrote to Kroger after the company announced it would introduce the technology, amid accusations that it was using facial recognition to show different customers different prices. In a letter of response obtained by Fast Company, Kroger defended the rollout, saying ESLs helped it manage the 1.3 billion price changes it implements each year and freed up associates to assist customers. Paula Walsh,Krogers director of retail operations, denied in the letter that the company was using facial recognition or collecting personal information from customers through the tags. Kroger dodged my questions but confirmed my key concerns: Its using electronic shelf labels to change grocery prices in real-time and collect data that could be used to jack up grocery prices for Americans, Warren tells Fast Company. Ill keep pushing to make sure consumers arent being exploited while they work hard to put food on the table. Wikander, for his part, dismisses the idea that retailers would use the technology that way. Just because you have the possibility of screwing your customers doesn’t mean that retailers will do that,” he says. “I don’t think retailers would typically do it, because the consumers are smarter than that. Wikander says it takes a typical business around a year or two and that, while the investment upfront is big, the labels last for many years. Indeed, for all the eeriness surrounding the labels, research shows that it might not be much of a change, price wise, for either consumers or businesses. Ioannis Stamatopoulos, a business professor at the University of Texas at Austin, says there is little evidence that digital shelf labels lead to significant price swings. He pointed to a 2025 study involving an American grocery store that found no evidence of the practice, and another involving an international grocery store that showed that prices tended to decline, particularly for items with short shelf lives. Much of his research, at least, suggests that the labels are most effective at stopping food waste, since it makes it easier for stores to offer sales on products like bananas and strawberries when theyre about to go bad. For now, the future of grocery shopping may look almost exactly like the pastexcept the price tag is oh-so-faintly glowing.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||