- Anthropic CEO Dario Amodei says there is a 25% chance AI could cause catastrophic harm.
- He also believes there is a 75% chance AI will bring enormous global benefits.
- Amodei has warned before about AI risks, including job losses and chip exports.
- His comments add weight to ongoing debates about AI regulation and safety.
When it comes to artificial intelligence, the conversation often swings between wild optimism and deep fear. But Anthropic CEO Dario Amodei recently gave a perspective that doesn’t neatly fit either side. Speaking at the Axios AI + DC Summit, Amodei openly estimated that there is a 25 percent chance that AI could lead to a catastrophic outcome for society.
It was not a casual slip of the tongue. Amodei described the possibility of “things going really, really badly,” outlining risks that range from destabilized social systems to large-scale misuse of AI, and even scenarios that could pose existential threats. In other words, he was not talking about funny text errors or everyday glitches, but risks so serious they could alter or even break the systems humanity relies on.
At the same time, Amodei made clear that his focus is on the other side of the equation: the 75 percent chance that artificial intelligence brings enormous benefits. He framed the future of AI as a balance of risk and reward, one that requires careful management rather than fear-driven paralysis.
Why a 25 Percent Chance of Doom Feels Different
On paper, a 25 percent risk might not sound like the end of the world. In fact, in other areas of life, one-in-four odds can even feel favorable. Casino players would gladly take such chances compared to the house edge in most games. But in the context of technologies that could reshape or destroy societies, those numbers look very different.
To make the idea more vivid, imagine if your car carried a 25 percent risk of exploding every time you turned the ignition. Most people would abandon the vehicle entirely. Amodei, however, suggested that his approach would be different: he would take apart the car, examine every piece, and see if it could be made safe to drive. That is the essence of his view on AI. Rather than retreat, he wants to actively shape it.
The CEO Who is Building the Risk He Sees
Amodei’s perspective carries unusual weight because of his position. He is not a commentator watching from the sidelines. As the CEO of Anthropic, the company behind the Claude AI assistant, he is one of the figures directly responsible for advancing the very technology he warns could go wrong.
This dual role, builder and cautious predictor, makes his stance stand out. Amodei has warned before about other dangers of artificial intelligence. He has suggested that AI could wipe out up to half of entry-level white-collar jobs. He has also spoken against the export of advanced AI chips to China, highlighting the potential for geopolitical consequences.
His willingness to publicly assign a number to AI’s risk of catastrophe adds another layer to the ongoing debate. Unlike vague comments about “uncertainty,” the 25 percent figure gives the public and policymakers something tangible to consider.
A Future of Enormous Promise
While Amodei did not sugarcoat the dangers, his outlook is not dominated by fear. He is also deeply convinced that artificial intelligence can bring extraordinary benefits if developed responsibly. His “75 percent chance things go really well” is not simply about optimism—it reflects the potential for AI to make breakthroughs in areas that matter to everyone.
The possibilities he alluded to include better healthcare systems, more efficient manufacturing, and even solutions to global crises such as climate change. With the right design and safeguards, AI could accelerate discoveries, streamline production, and open up new tools for addressing humanity’s toughest problems.
This is the tension at the center of his remarks. On one side, a future in which AI reshapes medicine, education, and the economy for the better. On the other, a future where AI runs out of human control or is misused in ways that spiral into disaster.
The Growing Public and Policy Debate
Amodei’s comments arrive at a time when public and government discussions about AI are intensifying. Policymakers in the United States and abroad are considering new rules for the industry. Concerns about safety, misuse, and energy consumption are surfacing alongside enthusiasm for innovation.
Other leaders in the field, including OpenAI’s CEO Sam Altman, have also expressed unease about the path AI could take. Altman himself has spoken about fears that people might end up letting AI dictate how they live their lives, which he described as both “bad and dangerous.”
These overlapping warnings are helping shape how lawmakers and the public view artificial intelligence—not just as a tool, but as a force with stakes that could reach into the very structure of society.
Amodei’s framing, however, is particularly striking. He did not just acknowledge the risks or highlight the benefits. By placing numbers on the possible outcomes, he gave the conversation a sharper edge.
The idea that the future might be “75 percent brilliant and 25 percent broken” is not a science fiction plot. It is the reality that one of the most influential figures in the industry believes we are facing.
Follow TechBSB For More Updates
