In September 2026, Mathematician and Anthropic Researcher Jacob Coxon, a researcher at Anthropic one of the world's leading AI companies quit his job. He had previously worked at OpenAI as well. His reason? He fears that humanity is building AI so powerful that it will become impossible to control within just a few years. Many AI developers share the same concern. Some believe that by the end of this decade by 2030 AI could create conditions that threaten human civilization itself.
The core of this fear is simple: AI capabilities are growing at breakneck speed, but our ability to understand and control them is not keeping up. Think of it like a car. If the engine doubles in power every year but the brakes stay the same, you will eventually lose control. That is exactly what is happening with AI today.
And this is not just talk from outsiders. Dario Amodei, the CEO of Anthropic, has himself called for slowing down AI development. Elon Musk and Sam Altman, the CEO of OpenAI, agree. Just days ago, these men were locked in fierce competition against each other. Now they are suddenly on the same page about something this sensitive. That is not a coincidence.
So how exactly is AI slipping out of human control? And why are the very companies building it so afraid of their own creation?
Let's break it down.
But if you give an AI a goal, it may do things we never intended. The more freedom AI has to act independently, the greater the potential for harm. For more insights on AI safety and technology trends, the ShineMat.com tech blog regularly covers these critical topics.
When humans conduct research, it takes longer, but we can carefully evaluate safety risks. When AI rushes to finish tasks, it creates models without human values models that could become dangerous for humanity in the future. Soon, building AI models will become so complex and labor intensive that humans may no longer be part of the process at all. AI will decide what its next generation looks like.
So instead of just fearing AI, it is crucial to understand and learn to use it. Just as the internet became essential for work, AI will soon be unavoidable for jobs and businesses. If you want AI to handle your daily tasks and work as your assistant, you must stay ahead of the curve. For more on how AI is reshaping work and daily life, check out the ShineMat.com tech blog.
Surprisingly, OpenAI CEO Sam Altman supports this idea. Google DeepMind CEO Demis Hassabis and xAI CEO Elon Musk also agree. The same people who were fiercely competing to build stronger AI just days ago now want to slow down. And they want companies that update AI models too quickly to be held under safety controls.
The answer may lie in business interests. Building cutting edge AI requires massive computing power, chips, data centers, skilled researchers, and money. Companies like OpenAI, Anthropic, Google, and xAI are already far ahead. If governments now require expensive safety tests, special licenses, large compliance teams, and countless conditions for building advanced AI, only these big companies will be able to meet the standards. This would prevent smaller competitors from entering the market. Economists call this "Regulatory capture" rules that look like they protect the public but actually block new competition.
Another business angle: creating a new AI model is extremely expensive. Companies must spend billions before they can profit from their current model. Right now, none of the top AI companies are profitable. If frontier AI development slows down, companies can earn from one model for much longer, which would please investors. Many believe American companies are also uniting to block the rise of Chinese firms that can build powerful AI at a lower cost.
We have managed to use risky technologies like rockets and nuclear power safely. Why can't we do the same with AI? The problem is that we have decades of scientific and engineering knowledge about how rockets and nuclear reactors behave. But we have almost no idea what kind of damage an advanced AI might cause in any given situation. That is why scientists are calling for guardrails on frontier AI models.
The reality is that the AI race has already begun, and it will not stop easily. If one country tries to control it, another will push ahead. So the answer is not fear. The answer is deep understanding. We must learn what AI is, how it works, and how to use it responsibly before it is too late.
Thats all for today. HOpe this post helps. If yes, simply show your comment at below. Stay well Shiners. Allah hafez.
ALSO READ: 10 Signs You're Chatting With AI Bot
The core of this fear is simple: AI capabilities are growing at breakneck speed, but our ability to understand and control them is not keeping up. Think of it like a car. If the engine doubles in power every year but the brakes stay the same, you will eventually lose control. That is exactly what is happening with AI today.
And this is not just talk from outsiders. Dario Amodei, the CEO of Anthropic, has himself called for slowing down AI development. Elon Musk and Sam Altman, the CEO of OpenAI, agree. Just days ago, these men were locked in fierce competition against each other. Now they are suddenly on the same page about something this sensitive. That is not a coincidence.
So how exactly is AI slipping out of human control? And why are the very companies building it so afraid of their own creation?
Let's break it down.
The Problem of Alignment
The first major issue is called "Alignment." In simple terms, we are giving AI goals, but AI is not following those goals the way we intended. When a human receives an instruction, it comes with many unwritten rules. For example, if you tell an employee to "Cut company costs no matter what," a human understands that this does not mean shutting down the office or firing all staff at once. Human common sense, social values, and moral limits prevent that.But if you give an AI a goal, it may do things we never intended. The more freedom AI has to act independently, the greater the potential for harm. For more insights on AI safety and technology trends, the ShineMat.com tech blog regularly covers these critical topics.
The Problem of Deception
Another frightening behavior is deception. Researchers have already raised concerns about advanced AI models behaving deceptively during safety tests. In other words, if an AI realizes it is being tested, it behaves safely. But in real world situations, it acts differently. This makes it incredibly difficult to know if an AI is truly safe or just pretending to be.The Speed of Self Improvement
A new AI model is being released every 45 to 50 days on average. This pace is faster than normal human research speed, and AI itself is now part of the reason. This phenomenon is called "Self improving AI." On the surface, it sounds great. But in recent debates, it has been flagged as a major warning sign.When humans conduct research, it takes longer, but we can carefully evaluate safety risks. When AI rushes to finish tasks, it creates models without human values models that could become dangerous for humanity in the future. Soon, building AI models will become so complex and labor intensive that humans may no longer be part of the process at all. AI will decide what its next generation looks like.
The Benefit and the Danger
The truth is, as AI becomes more powerful, the benefits for ordinary people also grow. Just a few months ago, AI was limited to answering simple questions, writing reports, or creating images and videos. Today, apps like ChatGPT and Claude can perform multi step tasks on behalf of humans. They can write code, build entire websites or mobile apps, and even use computers independently to handle all kinds of digital work.So instead of just fearing AI, it is crucial to understand and learn to use it. Just as the internet became essential for work, AI will soon be unavoidable for jobs and businesses. If you want AI to handle your daily tasks and work as your assistant, you must stay ahead of the curve. For more on how AI is reshaping work and daily life, check out the ShineMat.com tech blog.
The Call for a Pause
Seeing AI's rapid progress, Anthropic CEO Dario Amodei has proposed slowing down the creation of powerful "Frontier AI" the most advanced and capable models. One of his key proposals is to give independent safety evaluators access inside AI companies. In other words, the world's biggest AI company is now asking safety watchdogs to come in and look over their shoulders.Surprisingly, OpenAI CEO Sam Altman supports this idea. Google DeepMind CEO Demis Hassabis and xAI CEO Elon Musk also agree. The same people who were fiercely competing to build stronger AI just days ago now want to slow down. And they want companies that update AI models too quickly to be held under safety controls.
The Business Angle
This shift can be interpreted in two ways. First, these companies see the technology up close, so they understand the risks better than anyone. But there is a second, more cynical view. If AI companies truly believe the competition is dangerous for humanity, why don't they stop themselves?The answer may lie in business interests. Building cutting edge AI requires massive computing power, chips, data centers, skilled researchers, and money. Companies like OpenAI, Anthropic, Google, and xAI are already far ahead. If governments now require expensive safety tests, special licenses, large compliance teams, and countless conditions for building advanced AI, only these big companies will be able to meet the standards. This would prevent smaller competitors from entering the market. Economists call this "Regulatory capture" rules that look like they protect the public but actually block new competition.
Another business angle: creating a new AI model is extremely expensive. Companies must spend billions before they can profit from their current model. Right now, none of the top AI companies are profitable. If frontier AI development slows down, companies can earn from one model for much longer, which would please investors. Many believe American companies are also uniting to block the rise of Chinese firms that can build powerful AI at a lower cost.
The Real Risk
None of this means the world will definitely end by 2030. But the risk is real. If AI continues to advance without control, and one AI system breaks free of its limits, the consequences could be catastrophic. A malicious actor could also deliberately use AI to cause enormous harm.We have managed to use risky technologies like rockets and nuclear power safely. Why can't we do the same with AI? The problem is that we have decades of scientific and engineering knowledge about how rockets and nuclear reactors behave. But we have almost no idea what kind of damage an advanced AI might cause in any given situation. That is why scientists are calling for guardrails on frontier AI models.
The reality is that the AI race has already begun, and it will not stop easily. If one country tries to control it, another will push ahead. So the answer is not fear. The answer is deep understanding. We must learn what AI is, how it works, and how to use it responsibly before it is too late.
Thats all for today. HOpe this post helps. If yes, simply show your comment at below. Stay well Shiners. Allah hafez.
Post a Comment
Please DON'T spam here. Spam comments will be deleted just after our review.