To be clear, you don't need AI for this.

You can also just call the railroad and report the bridge as damaged.

Hoaxes and pranks and fake threats have been around forever.

I love hoaxes. But this also neglects the social and viral aspect. If lets say if an aged local member for parliament sees an image of the bridge after coming back from the pub, he will call into the authorities responsible. Now you can think of many other people who, upon seeing this, and having an initial reaction to it, have the power to enforce an action.

Calling directly into the railroad bypasses an authority chain. It negates the virality of it. These viral images are viral because they get shared and spread on their own just like a virus.

Telephone calls into authorities were never viral, they could never be spread. Although they may well have caused the desired reaction without spreading first! Many hoaxes back in the day were somewhat viral and did get spread, but the hoax went to the newspapers or the community first and spread there. A well crafted press release, some additional letters to the traditional media etc. A believable image makes for more believability. The hoax got spread because it was hard to debunk it as it was distributed before the debunking. Bypassing the effort to spread the hoax removes chances of effects.

Edits: my initial thought was "no trains run after midnight anyhow" as except on a few main lines its hard to find trains in the UK at night - so the cost of the bridge closure may have been very small. That with the amount and quality of the staff operating at that time of night. Taken together this leads to less of a cost of reaction, more of a chance of a knee jerk reaction from staff, less ability to consult nearby awake engineers and survey damage IRL. So while the hoaxers cannot plan an earthquake(!) it probably wouldn't have succeeded if the earthquake happened at 11am.

Again, I see this argument.

“Bad X has happened before and unsolved. Why worry about bad X^2?”

Personally I’d prefer if it remained at X so solutions can catch up. But that’s just me.

Any hoax can be easily fought, if the punishment of getting caught is severe enough.

The problem is the justice system, that is optimized to protect a criminal and to offload the costs to the society, which is happy to be distracted with identity and moral supemacy arguments.

That leaves much more of a paper trail. People routinely are fined and jailed for pulling off such "pranks", partially because "fake threats"/"abuse of emergency response resources" are an exception to many freedom-of-speech laws.

A fake photo of a collapsed bridge however won't cross that criminal threshold.

You also don't need gunpowder to kill someone with projectiles, but gunpowder changed things in important ways. All I ever see are the most specious knee-jerk defenses of AI that immediately fall apart.