- The anthropogenic CEO Dario Amodei estimates a chance of 25% of AI leads to the disaster
- He always thinks that AI is worth investing and the benefits prevail over the risks
- His comments correspond to increasing public and political conversations on AI risks and regulations
One in four ratings may seem fairly good in certain circumstances. It’s much better than most casino games, for example. However, it is apparently quite unlikely that the anthropogenic CEO Dario Amodei seems indifferent after having struck the chance that artificial intelligence leads to an end -of -company disaster at 25%.
“I think there are 25% of things that things will really happen, really bad,” said amodei cheerfully at the Axios Ai + DC summit when he asked him about his (p) doom – probability of condemnation – belief around AI. But it is more focused on the “75% chance that things are really going on, very good”.
By “really, really bad”, he does not mean that your phone corrects the “duck” to something worse. It means scenarios large enough to threaten societal systems, existential risks, INF results.
. @Jimvandehei asks the CEO of @anthropic @darioamodei what probability it will give that AI ends with a disaster: “I think there is 25% likely that things are really going, really bad.” #Axiosaisummit pic.twitter.com/9d7eqldyncSeptember 17, 2025
For an industry often soaked in the utopian promise or reduced to the fear of science fiction, the attitude of amodei both in the chances of an apocalypse and why it always advances with technology.
Amodei is not the only one to express discomfort, but it is in a rare position. As the CEO of the company behind Claude, he is not a passive observer. It shapes the trajectory of this technology in real time. His team builds the very systems whose potential and the danger he weighs.
If someone told you that there was a chance of 1 in 4 explode each time you turn the key, you could start walking more. Amodei would apparently become a mechanic and check the car before entering.
Ai doom
Nor is it the only warning issued by Amodei on AI. He has already warned that AI could eliminate half of all the work of entry -level collar and sounded the alarm against American exports of high -end fleas to China. This is what makes the frame of amodei so useful. He recognizes the risk, quantifies uncertainty, but leaves room for the agency.
On the other hand, the “75% chances of amodei go very well” are not optimistic for itself. This implies the conviction that AI could produce enormous advantages for everyone. This could lead to an improvement in drugs, a more effective manufacturing and perhaps even strategies to combat existential crises such as climate change (although a key element to solve this problem could be the energy required for AI models to work).
But the risk of 25% requires that these advantages be constructed carefully, in consideration for safety measures and regulations. Because if the future is 75% brilliant and 25% broken, the question is: what are we going to do to keep the weight on the right side?
