The failed ones were known to be dangerous. Mostly that big problem was kept under radar of the public, oversight failed big time or was complicit. The other big problem is, we keep running old rustbuckets and dubious designs. More than half of the worlds reactors should probably be shut down immediately, and most of the rest is also due for a replacement very very soon.

How do we know that this time is different?

See, the thing that goes wrong doesn’t have to be even a technological one. Fission is a process that spontaneously can go wrong unless harnessed by systems that are built by thousands of people. Anything can go wrong and anyone can screw up. All can happen and the same time as demonstrated by the previous disasters.

You say that the current reactors are very dangerous and should be closed down. Please give me a material by the plant builders and operators that communicates these dangers with the people that might get exposed to a failure.

Fission can spontaneously go wrong in the same way all the protons in your body could just decide to decay. A reactor is not magic, just physics.

Your narrative sounds like you're claiming that we didn't learn from Chernobyl and that Fukushima is proof of that. That's not true, no one is going around claiming that ancient designs are somehow becoming safer and safer. Fukushima was constructed literally before Chernobyl was. Back then generation II reactors took an active approach to safety. Operators might have to take action and power to the plant needed to be always available in order to safely manage some faults. They took the approach of "lets get five redundant diesel generators and have triple redundant control cabling, sensors, and valves" whereas newer Generation IV reactors are passively safe when things go wrong. Even on the AP1000 which was approved by the NRC back in 2005 is passively safe even in a plantwide blackout with a gravity based evaporative cooling system that can last 72 hours before needing an additional source of water to fill it back up. Around 72 hours after shutdown the decay heat should be able to boil a little over a gallon per second of 30C water. All things considered it's pretty trivial to get a gallon per second of water to the reactor within 3 days time. The new SMR from NuScale that just got approved is even better given that the pool the reactor sits in is sized large enough that by the time the water in the pool is gone it'll have reduced to the point where it's cold enough to just air cool the reactor so it's passively safe indefinitely.

Even in the case of Fukushima, I'd like to point out that all things considered, zero people died from acute radiation poisoning and IIRC only one case of lung cancer was likely attributed to exposure from Fukushima. Fukushima was basically a worst case scenario for any Gen II reactor. The only reason Chernobyl was so much worse was because of the incredibly unsafe design out of the USSR. They didn't care, they didn't bother with a containment building at all, had operators blindly following orders from management even more incompetent than the operators were, operated the reactor way outside of the limits that were placed on the reactor by removing all the control rods which was never supposed to be done, etc. You can pin Fukushima on the Nuclear industry, Chernobyl was solely because of the callous indifference and incompetence of the USSR.

>"And what would be the excuse when the new reactors fail disastrously? "

CANDU reactors are failsafe and have been for 50 years. Fukushima or Chernobyl style disaster in phisically impossible with their heavy-water, unpressurised design.