Mouz

Is AI making us dumber?

· Business Insider

Visit lej.life for more information.

My family is in its natural state when we're in debate, not just about what should happen, but what has happened.

Growing up, before my grandparents gave in and wired their home for the new millennium, their full set of encyclopedias proved the only way to settle disputes. Many arguments, particularly over any event that occurred after the encyclopedias were printed, went unsolved. Then came Google's ubiquity. All of this access to information didn't just close the chasm between what we could know in seconds and what we would have to ponder for months, it changed how the world remembers.

Researchers called this the "Google effect." They found people recalled where to find specific information better than they remembered the information itself when they knew they could easily find it again. "We are becoming symbiotic with our computer tools, growing into interconnected systems that remember less by knowing information than by knowing where the information can be found," the researchers wrote in 2011. Some worried that cognitive offloading to Google was "making us stupid," a possibility raised by an Atlantic cover story. Others argued Google was democratizing access to information and let us trade hours spent scouring library stands for supercharged thinking.

You don't need to Google any of this to know why it sounds so familiar. Early research into how generative AI affects our brains has resurfaced the same talking points: overreliance on AI will weaken mental persistence, flatten creativity, atrophy our critical thinking skills, and degrade our relationships. Experts in machine learning, creativity, social behavior I spoke to said we can glean some insight from the fallout of past innovations, but the totalizing pervasiveness of AI is unparalleled.

AI could pose a bigger risk to our brains than past innovations because "the tool is completely different in nature," says Nataliya Kosmyna, a researcher at MIT who published one of the most widely cited pieces of research on AI and cognitive decline last year, showing that people who had access to gen AI for writing essays performed worse over time than those who used Google or had no aid. Kosmyna says the widely circulated comparisons of AI to a calculator, which has also been used by Sam Altman, is a fallacy. "You don't fall asleep and wake up with a calculator. You don't talk to the calculator about everything you have in your mind."

If AI integrates into every facet of our lives, as its creators predict, it's going to change how we think. What we don't know is how permanent those changes may be.

Innovation has always evoked fears that our brains will be rewired for the worse. Socrates worried that the written word would dilute people's memories. Some posited that the telegram would be an end to poetry. The calculator was going to atrophy our mental math skills. None of these things happened in the cataclysmic sense. But they did happen, gradually, not from any one technology itself but from waves changes over decades, like the proliferation of smart phones, more lenient grading policies, and school funding shortages. Fewer than 10% of people read poetry, and the art form has dropped from its 1800s heyday when books of popular poetry regularly sold out. The percent of students scoring at least basic proficiency in math have been dropping from their 2013 high, according to test results of 12th grades from the National Assessment of Education Progress. Writing became one of the first examples of cognitive offloading, or, making records so we don't have to remember everything.

AI, however, is less an evolution and more a big bang. Experienced workers are wondering what tasks they can offload to AI, while educators and employers worry that young people might miss the chance to develop skills. There's a difference between spending decades coding and then seeing how AI can enhance one's work, and only learning to vibe code. Researchers at the Wharton School of the University of Pennsylvania gave 1,000 high school students in Turkey gen AI math tutors; one mimicked a standard ChatGPT, the other had more guardrails, and instead of generating direct answers, would give hints, and also provided information specific to the problems that came from teachers, including correct solutions to practice problems, and explanations of common mistakes. Students performed better with both tools, but when taken away, those who had used the more standard ChatGPT-like tool did worse than those who never had access to AI for the work.

There's never been a tool like this that has offered us the opportunity to let thinking be done for us.Nataliya Kosmyna

Something similar happened when GPS became everyone's personal navigator. A 2020 study from researchers at McGill University found that the longer people relied on GPS, the worse spatial memory they had when they needed to navigate without it. When the researchers followed up with a small sample size three years later, they found that increased GPS use had led to even steeper decline in spatial memory.

This doesn't mean those who made it through school and their early careers before AI are spared. Creativity and cognition take continued practice. Take pilots. Research from 1971 found that pilots can maintain hand-eye coordination skills like scanning instruments and handling flight controls well, even after they don't fly for four months. But the pilots' cognitive skills declined in the same period, as they struggled to recall and track the necessary procedural steps, visualize the plane's position, and do mental calculations.

If people never learn certain skills to begin with, we could reach a point where these skills atrophy and fade from society, or go the way of hobbies. "There is a high risk if young people never learned this critical thinking, because they have the convenience of an AI thinking for them, they might not be able to develop that," says Michael Gerlich, head of the Center for Strategic Corporate Foresight and Sustainability at the Swiss Business School.

When there's a convenient out, AI can make us less resilient in the face of challenges. In an April 2026 preprint study, researchers gave participants math problems with fractions. One group used an AI assistant that could provide the answer with minimal prompting for 12 of the questions, but had to answer the final three problems on their own. Those who had access to AI were more likely to solve the first 12 problems than the control group, who had no AI access for the 15 questions. But they also proved more less likely to correctly solve and more likely to skip the last three. The results suggest that people who used AI did not persist through the difficulty as often, and the change occurred after only using the help for 10 minutes. "A mentor or companion doesn't just answer questions, but also scaffolds learning, tracks progress, and prioritizes the other person's growth over immediate results," the authors wrote. "In contrast, current AI systems are fundamentally short-sighted collaborators — optimized for providing instant and complete responses, without ever saying no."

The findings are alarming, but Grace Liu, a machine learning PhD candidate at Carnegie Mellon University who authored the study, says we don't yet know what this small study says about ways our brains would change after lots of AI use. "Ten minutes of use definitely wouldn't do some long-term brain or cognitive decline," she says. "It's an open question as to what happens cognitively if we use it over and over again for extended periods of time, and that's something that we would need to do longitudinal studies to investigate."

Researchers from Georgetown University have investigated how the emergence of ChatGPT affects creativity. They analyzed more than 370,000 college application personal essays from before and after ChatGPT's rise to popularity, finding that human-written essays had more new ideas than AI-assisted essays, but that AI essays had more unique language. While AI can lead to more creative language use in a single essay, its use can also dull creativity across a group. Adam Green, director of the Georgetown Laboratory for Relational Cognition, who worked on the essay research, says AI is different from past shortcuts, because it's the first technology to ideate for us. "Whereas Google helps me find the thing that I thought of finding, AI thinks of finding it," he says.

It's much too soon for robust data on AI's cumulative effects on our brains over time, and the experts I spoke to emphasized a need for more and longer studies (the research is so new, much of it has yet to be peer-reviewed). Even those who already know how to do things, like write well, code, or do complex math, can still lose ability over time. "Any skill that we acquired as humans does need to be retained and retrained," Kosmyna says.

The studies published so far are small. But each raises a small alarm as we're living through the experiment. The experts I spoke to expose their brains to AI with caution. The best practice might be emphasizing awareness "of what the cognitive effects are and so that we can choose to prioritize the skills that we want to keep independently, and then the skills that we're OK with outsourcing and offloading," Liu says. Green says we must honor the blank page. "What I'm really worried about is the creative thinking, the creative intelligence, that we develop through practice," he says. "There's never been a tool like this that has offered us the opportunity to let thinking be done for us." Kosmyna says she "proudly" does not use large language models for her personal life, and keeps the AI tools she builds tied to her research.

Despite the optimism around Google, we didn't get smarter or freer for deep work. Boundaries between work and home blurred, and many now work long hours since the advent of the internet. IQ scores rose throughout the 20th century, by about an average three points per decade. But people are testing lower in several categories between 2006 and 2018, with the steepest drops among those ages 18 to 22 — the most digital native adults in the pool. Our attention spans have shrunk, with people flitting back and forth between different content online. But researchers suggest this is a habit change rather than a neurological one — we can retrain ourselves to focus longer, but that may only happen if we remove the distractions we've now built into our lives.

What we know about AI doesn't neatly tie itself into something as succinct as the Google effect. That's indicative of the sprawling way AI touches our lives. But we do know that guarding creativity, critical thinking, and persistence alongside it will become our next challenge.

Amanda Hoover is a senior correspondent at Business Insider covering the tech industry. She writes about the biggest tech companies and trends.

Read the original article on Business Insider

Read full story at source

China May Restrict Access to Its Most Powerful AI Models

· Time

The ByteDance and Alibaba apps arranged on a smartphone. —Lam Yik—Bloomberg/Getty Images

Chinese AI companies have made inroads globally by giving their models away for free. Now Beijing is weighing whether to stop them. Chinese authorities have held talks with Alibaba, ByteDance, and Z.ai about whether to restrict foreign access to their most advanced models, including ones not yet released, Reuters reported today, citing three people familiar with the discussions. 

Nothing has been decided yet, and the ministries involved have made no official comment, but officials have gone as far as sketching options—including a bar on public release or a limit to domestic use only.

Visit esporist.com for more information.

That would be a dramatic U-turn. Most Chinese AI companies currently release their models "open-weight"—publishing the weights online so anyone can download the system and run it themselves. That openness is precisely how they have made inroads globally despite lagging America's best models by seven months on average. The models are free, and some U.S. businesses have adopted them to cut costs compared with proprietary American models.

Closing access would mean surrendering the very lever that has driven China's rise. A trailing player doesn't abandon its biggest advantage unless the concern is national security. In that sense, Beijing would be following Washington, which has already restricted Anthropic's Mythos and OpenAI's GPT-5.6 over concerns about the models' ability to find software vulnerabilities.

"China will need to reckon with the reality that models that reach certain capabilities are unsafe. It is going to have the same conversations the White House has had over the last many months," says Scott Singer, a fellow in the technology and international affairs program at the Carnegie Endowment for International Peace. "China is going to have to balance the benefits of access to global markets with a desire to control a technology that is central for national security."

Late last month, Z.ai released GLM 5.2, which it claimed had matched Mythos' ability to spot bugs. Once a model’s weights are published online, they are impossible to recall or add safeguards, which presents additional security challenges compared with proprietary models. It also means any forthcoming restrictions would only impact future models.

It follows an escalating tit-for-tat between the two countries. Chinese tech giant Alibaba banned Claude Code internally and asked employees to remove Claude from their work computers, the Information reported Friday. Days earlier, a developer had discovered that Anthropic quietly slipped code into Claude Code that tried to work out whether the person using it was in China or connected to a Chinese AI lab, reading signals like their time zone and network address. An Anthropic engineer, responding on X, said the code was added in March to fight a copying technique called "distillation," that stronger safeguards were now in place, and that it would be removed the next day.

“Distillation,” or using the outputs of a smarter AI model to improve the performance of a weaker one, has become a point of controversy in the AI race. Chinese AI models trail America’s best by roughly seven months on key benchmarks. To catch up, Anthropic and others say Chinese labs are distilling their models—a violation of their terms. Anthropic published a report in February claiming other Chinese firms DeepSeek, Moonshot, and MiniMax generated 16 million exchanges with Claude through approximately 24,000 fraudulent accounts. In June, Anthropic reportedly sent a letter to U.S. officials accusing Alibaba of “brazenly” attempting to distill Claude’s capabilities. The hidden code that rattled Alibaba was built to help catch exactly this.

Anthropic frames distillation as a national-security threat, warning that foreign labs could funnel stripped-down capabilities into military and surveillance systems and let authoritarian governments "deploy frontier AI for offensive cyber operations, disinformation campaigns, and mass surveillance." There's a commercial angle too. If Chinese labs siphon American capabilities and release them open-source, they erode the very business model those labs are banking on.

Anthropic has argued since February that no single firm can solve this—hence its lobbying for a coordinated front of industry, cloud providers, and government. With Washington already deciding who may use Claude Mythos and OpenAI’s GPT-5.6, that front no longer looks far-fetched.

Read full story at source

CNN star offers stunning update on Mitch McConnell as rumors swirl about former Senate leader’s health

· The Independent