With AI safety debates making headlines and weather warnings dominating the news, you might wonder: what about the safety of people already locked away? The answer is more surprising than you'd expect.

Prison staff are dramatically safer today than they were 80 years ago. In 1942, there were 182 assaults on staff per 1,000 prisoners — nearly one in five inmates attacked a guard each year. By 1998, that figure had crashed to just 44 per 1,000, a staggering 76% drop.

This isn't just a statistical curiosity. It represents thousands of prison officers going home without injuries, careers not cut short by violence, families not getting phone calls from hospital wards. In 1942, working in a British prison meant facing assault was almost inevitable. Today, while still dangerous, it's far less likely.

The transformation didn't happen overnight. The steepest decline came in the post-war decades, as Britain rebuilt not just its cities but its institutions. Better training, improved conditions, and reformed practices turned prisons from powder kegs into more controlled environments.

But here's what the long-term trend masks: this data stops at 1998. We're missing 26 years of the story. Prison populations have surged since then, overcrowding has worsened, and staff shortages have become chronic. Without current figures, we're flying blind on whether this remarkable safety improvement held.

The 1942 baseline tells us something profound about institutional change. When Britain was fighting for its survival abroad, its prisons were battlegrounds too. Prison staff faced violence rates that would be considered a crisis today. Yet somehow, without the technology or resources we take for granted, the system transformed itself.

What changed? Partly, society's understanding of what prisons should achieve. The post-war consensus moved away from pure punishment toward rehabilitation. Staff became more than just turnkeys — they became part of a system designed to return people to society, not simply warehouse them.

Today, as politicians promise to build more prisons and courts hand down longer sentences, this 56-year decline in staff assaults offers a lesson. Safety behind bars isn't just about hiring more guards or building higher walls. It's about creating environments where violence isn't the default response to conflict.

The missing question isn't whether AI will be safe, or whether the weather will improve. It's whether the remarkable transformation of prison safety that took place between 1942 and 1998 survived the overcrowding crisis that followed. Without that data, we're debating the future of criminal justice while ignoring whether we've kept the progress we already made.