Latest news with #MaxPlanckInstitute
Tatler Asia
5 days ago
- Tatler Asia
Do you sound like ChatGPT? Research says AI is changing how we speak
Stop mid-sentence and listen to yourself talk. Notice how you 'delve' into topics or describe things as 'meticulous'? Congratulations—you've been ChatGPT-ified Research from the Max Planck Institute has revealed a startling linguistic shift: since ChatGPT debuted 18 months ago, humans have increased their usage of AI-favoured vocabulary by 51 per cent. Words like 'adept', 'realm' and 'navigate', which are commonly used by ChatGPT, are suddenly everywhere, from boardroom presentations to dating app messages. We're not just using artificial intelligence—it turns out AI is shaping how we speak. This isn't accidental. Every time we interact with AI-generated content, while AI absorbs our patterns, we, as humans, absorb its patterns too. AI-speak sounds confident and professional, and we reach for its linguistic safety net. Also read: Can ChatGPT understand who we are? But here's the problem: uniformity kills personality. As we streamline our speech through ChatGPT filters, we iron out regional dialects, personal quirkiness and grammatical stumbles that make our speech uniquely our own. Consider the difference between 'I enjoy navigating through multicultural spaces' and 'I love travelling'. One sounds impressive and professional, but the other sounds like an actual person, someone relatable. There's no denying that AI is here to stay. And AI is shaping our culture, including our communication style. But it's up to us on how we strike a balance between what is beautifully, chaotically human and the confident 'perfectly sophisticated' voice that AI brings on board.
Yahoo
5 days ago
- Science
- Yahoo
Nobel Prize Winner Warns About Astronomers Using AI to Make Discoveries
A team of astronomers say they've gleaned the mysterious traits of our galaxy's black hole by probing it with an AI model. But a pretty big name on the field is throwing a little bit of cold water on their work. Just a little bit. Reinhard Genzel, a Nobel laureate and an astrophysicist at the Max Planck Institute, expressed some skepticism regarding the team's use of AI, and the quality of the data they fed into the model. "I'm very sympathetic and interested in what they're doing," Genzel told Live Science. "But artificial intelligence is not a miracle cure." Raging at the center of the Milky Way some 26,000 light years away is Sagittarius A*, a supermassive black hole with over 4.3 million times the mass of the Sun, and an event horizon nearly 16 million miles in diameter. Back when it wasn't clear what Sagittarius A* was other than a weird bright object in the galactic center, Genzel and fellow astrophysicist Andrea Ghez illuminated its colossal scale and eventually proved that it was a supermassive black hole, a feat that earned them both a Nobel Prize in physics in 2020. But much of our galaxy's dark, beating heart remains a mystery, as do supermassive black holes in general. How and when do these cosmic behemoths form, and how do they gain such incredible mass? Astronomers agree that they would have to have been formed in the early universe, but the rest remains contentious. One reason is that no star is heavy enough to directly collapse into an object of a supermassive black hole's size. True, they can grow by swallowing nearby matter, like an unfortunate star that wanders too close, or even merging with another black hole, but that doesn't explain all cases. Some are so massive that the time it'd take for them to accrete enough matter to reach their observed size would be older than the universe itself. A breakthrough came in 2022, when astronomers revealed the first image of Sagittarius A* taken with the Event Horizon Telescope, three years after the same observatory — which is actually made up of several radio telescopes scattered across the globe — was used to stitch together humankind's first image of a black hole whatsoever. But the image — and the data that comprised it — was fuzzy. There wasn't enough detail present to tease out the black hole's structure or behavior. That's where this latest work, detailed in three studies published in the journal Astronomy & Astrophysics, comes in. In a nutshell, the astronomers trained a neural network on millions of synthetic simulations using discarded ETH data that was deemed too grainy to decode, largely due to the interference introduced by the Earth's atmosphere. Once the AI model cut its teeth on the synthetic data, it looked at the real observations of Sagittarius A* and produced a much clearer image. "It is very difficult to deal with data from the Event Horizon Telescope," coauthor of the main study Michael Janssen, an astrophysicist at Radboud University in the Netherlands, told Live Science. "A neural network is ideally suited to solve this problem." The AI-enhancement suggested that the supermassive black hole is rotating somewhere between 80 to 90 percent of its maximum possible velocity, which is blindingly fast, as these objects can spin at a significant fraction of the speed of light. Its rotation axis, in fact, appears to be pointing towards the Earth. The AI model also revealed that the black hole's emissions are coming from its accretion disk — the glowing disc of hot matter swirling just outside its event horizon — and not an energetic outburst called a jet that's produced by the black hole's absurdly powerful magnetic fields. Genzel isn't totally sold on these findings. The renowned astrophysicist told Live Science that the poor quality of the data fed into the model could have biased it in unexpected ways. As a result, what we're seeing could be heavily distorted, and shouldn't be taken at face value, he said. That isn't to say we should write off "AI" — a buzzword whose meaning has been diluted by tech companies shoving their unfinished products into every facet of modern life — being used in astronomy. With the number of telescopes now in operation, it's impossible for humans to sift through all the data they collect manually. The European Space Agency's monumental Gaia survey, for example, has cataloged over two billion stars and counting. AI is already being used to classify cosmic objects. But it's worth being cautious about its application in the field all the same, since scientists aren't immune to hype, either. Jannsen, for his part, seems well aware that his team's AI findings are far from definitive, but is sticking to his guns. "That we are defying the prevailing theory is of course exciting," Janssen said in a statement. "However, I see our AI and machine learning approach primarily as a first step. Next, we will improve and extend the associated models and simulations." More on space: Long-Dead NASA Satellite Suddenly Lets Out Epic Blast of Energy
Yahoo
5 days ago
- Science
- Yahoo
Losing Our Voice: The Human Cost of AI-Driven Language
Losing Our Voice: The Human Cost of AI-Driven Language originally appeared on L.A. Mag. Not only is AI influencing our written communication, but it is also altering our verbal interactions. The Max Planck Institute for Human Development in Berlin has conducted research on the impact of ChatGPT on human speech by analyzing 280,000 academic YouTube videos. Since the release of ChatGPT in 2022, certain words have become more prevalent in everyday vocabulary, with terms such as 'delve,' 'meticulous,' 'realm,' and 'adept' being used 51% more frequently. Researchers at the Max Planck Institute for Human Development confirmed that ChatGPT favors these words to a greater extent than humans, as seen in a previous study comparing human- to AI-edited texts. The word 'delve' has proven to be a front runner in determining how speakers 'internalize this virtual vocabulary into daily communication,' according to Hiromu Yakura, the study's lead author and a are unaware of the shift in their language, but what's more concerning is how we sound. Despite research being primarily focused on vocabulary, there are suspicions that AI is beginning to shape our tone, with people speaking with more structure and less emotional expression. AI, a tool that seemingly improves our grammatical and structured communication, creates a divide in our collaborative communication through the suspicion it arouses. Mor Naaman, professor of Information Science at Cornell Tech has outlined three levels of human detachment caused by the adoption of virtual vocabulary. The first level is the loss of authenticity and vulnerability in our speech that proves our availability as human beings. The second level involves lessened attention and effort in our communication, relating to our level of care. The final level consists of our ability to show our real selves: our humor, our competence, or our regret. Naaman uses an example of texting someone, 'I'm sorry you're upset' versus 'Hey, sorry I freaked at dinner, I probably shouldn't have skipped therapy this week.' The distinction between a structured apology versus a personal explanation makes all the difference between ChatGPT and a real human being. We are losing the autonomy of our speech and the trust in our communication with others, and soon, we may lose agency over our thinking. 'Instead of articulating our own thoughts, we articulate whatever AI helps us to articulate…we become more persuaded,' Naaman says. Ultimately, our personhood is at stake. ChatGPT creates stereotypes about communities and what 'correct' English is. But, instead of adopting a homogeneous vocabulary and a 'correct' method of communication, we must self-regulate and celebrate the imperfections that build trust and make us human. This story was originally reported by L.A. Mag on Jun 23, 2025, where it first appeared.
Arab News
22-06-2025
- Science
- Arab News
Early humans survived in a range of extreme environments before global migration, study says
WASHINGTON: Humans are the only animal that lives in virtually every possible environment, from rainforests to deserts to tundra. This adaptability is a skill that long predates the modern age. According to a new study published Wednesday in Nature, ancient Homo sapiens developed the flexibility to survive by finding food and other resources in a wide variety of difficult habitats before they dispersed from Africa about 50,000 years ago. 'Our superpower is that we are ecosystem generalists,' said Eleanor Scerri, an evolutionary archaeologist at the Max Planck Institute of Geoanthropology in Jena, Germany. Our species first evolved in Africa around 300,000 years ago. While prior fossil finds show some groups made early forays outside the continent, lasting human settlements in other parts of the world didn't happen until a series of migrations around 50,000 years ago. 'What was different about the circumstance of the migrations that succeeded — why were humans ready this time?' said study co-author Emily Hallett, an archaeologist at Loyola University Chicago. Earlier theories held that Stone Age humans might have made a single important technological advance or developed a new way of sharing information, but researchers haven't found evidence to back that up. This study took a different approach by looking at the trait of flexibility itself. The scientists assembled a database of archaeological sites showing human presence across Africa from 120,000 to 14,000 years ago. For each site, researchers modeled what the local climate would have been like during the time periods that ancient humans lived there. 'There was a really sharp change in the range of habitats that humans were using starting around 70,000 years ago,' Hallett said. 'We saw a really clear signal that humans were living in more challenging and more extreme environments.' While humans had long survived in savanna and forests, they shifted into everything from from dense rainforests to arid deserts in the period leading up to 50,000 years ago, developing what Hallett called an 'ecological flexibility that let them succeed.' While this leap in abilities is impressive, it's important not to assume that only Homo sapiens did it, said University of Bordeaux archaeologist William Banks, who was not involved in the research. Other groups of early human ancestors also left Africa and established long-term settlements elsewhere, including those that evolved into Europe's Neanderthals, he said. The new research helps explain why humans were ready to expand across the world way back when, he said, but it doesn't answer the lasting question of why only our species remains today.
Yahoo
21-06-2025
- Science
- Yahoo
Nobel laureate concerned about AI-generated image of black hole at the center of our galaxy
When you buy through links on our articles, Future and its syndication partners may earn a commission. The monster black hole at the center of our galaxy is spinning at ear "top speed," according to a new artificial intelligence (AI) model. The model, trained partially on complex telescope data that was previously considered too noisy to be useful, aims to create the most detailed black hole images ever. However, based on the questionable quality of the data, not all experts are convinced that the AI model is accurate. "I'm very sympathetic and interested in what they're doing," Reinhard Genzel, an astrophysicist at the Max Planck Institute for Extraterrestrial Physics in Germany and one of the winners of the 2020 Nobel Prize in physics, told Live Science. "But artificial intelligence is not a miracle cure." For decades, scientists have been trying to observe and characterize Sagittarius A*, the supermassive black hole at the heart of our galaxy. In May 2022, they unveiled the first-ever image of this enormous object, but there were still a number of questions, such as how it behaves. Now, an international team of scientists has attempted to harness the power of AI to glean more information about Sagittarius A* from data collected by the Event Horizon Telescope (EHT). Unlike some telescopes, the EHT doesn't reside in a single location. Rather, it is composed of several linked instruments scattered across the globe that work in tandem. The EHT uses long electromagnetic waves — up to a millimeter in length — to measure the radius of the photons surrounding a black hole. However, this technique, known as very long baseline interferometry, is very susceptible to interference from water vapor in Earth's atmosphere. This means it can be tough for researchers to make sense of the information the instruments collect. "It is very difficult to deal with data from the Event Horizon Telescope," Michael Janssen, an astrophysicist at Radboud University in the Netherlands and co-author of the study, told Live Science. "A neural network is ideally suited to solve this problem." Janssen and his team trained an AI model on EHT data that had been previously discarded for being too noisy. In other words, there was too much atmospheric static to decipher information using classical techniques. Through this AI technique, they generated a new image of Sagittarius A*'s structure, and their picture revealed some new features. For example, the black hole appears to be spinning at "almost top speed," the researchers said in a statement, and its rotational axis also seems to be pointing toward Earth. Their results were published this month in the journal Astronomy & Astrophysics. Pinpointing the rotational speed of Sagittarius A* would give scientists clues about how radiation behaves around supermassive black holes and offer insight into the stability of the disk of matter around it. RELATED STORIES — New view of the supermassive black hole at the heart of the Milky Way hints at an exciting hidden feature (image) — Sagittarius A* in pictures: The 1st photo of the Milky Way's monster black hole explained in images — The 1st Milky Way black hole image was groundbreaking — the next could be even better However, not everyone is convinced that the new AI is totally accurate. According to Genzel, the relatively low quality of the data going into the model could have biased it in unexpected ways. As a result, the new image may be somewhat distorted, he said, and shouldn't be taken at face value. In the future, Janssen and his team plan to apply their technique to the latest EHT data and measure it against real-world results. They hope this analysis will help to refine the model and improve future simulations. This story was provided by a sister site of
