There actions are still not without fault, but I can imagine a few more important aspects of this issue that we could be debating.

"Highly Dangerous" seems excessive

In the case of a bullet fired at a precisely vertical angle (something extremely difficult for a human being to duplicate), the bullet would tumble, lose its spin, and fall at a much slower speed due to terminal velocity and is therefore rendered less than lethal on impact. However, if a bullet is fired upward at a non-vertical angle (a far more probable possibility), it will maintain its spin and will reach a high enough speed to be lethal on impact. Because of this potentiality, firing a gun into the air is illegal in most states, and even in the states that it is legal, it is not recommended by the police. Also the MythBusters were able to identify two people who had been injured by falling bullets, one of them fatally injured. To date, this is the only myth to receive all three ratings at the same time.

A bullet will lose sufficient velocity so that it no longer lethal relatively quickly. As an example we can use FBI specs for maximum range of the NATO standard issue M855 cartridge for 5.56mm , which is 3600 yards. Ballistics show a trajectory with 2200 foot drop at that range and a final velocity of 490fps with a retained energy of only 33ft/lbs (about 3% of muzzle energy). To attain impact at 3600 yards requires the rifle to be elevated at about 11 degrees. In reality, the bullet is likely to lose enough rotational velocity that it will destabilize much earlier, which will even more rapidly decrease velocity and power.

So, put simply, for this cartridge shooting in the air at an angle greater than 11 degrees relative to the horizon would not be particularly dangerous.

The roof was constructed of corrugated aluminum. Not the strongest stuff in the world, but the bullet passed through two sheets of this material on its way in. We estimated the entry angle at around 40-45° by looking upward through the holes, aligning the light that shined through them (the following day).

Having seen it first hand, I would dispute the assertion that shooting in to the air at a high angle isn't particularly dangerous.

There are certainly cartridges in production that carry dramatically more kinetic energy than the example I used, but the more powerful cartridges tend to be more rare.

Note that as velocity bleeds off the trajectory degrades faster and faster, resembling an upside down exponential curve. The angle of impact is much greater than the angle at which the round was fired; more so as distance increases.

The holes in the aluminum roof were round (ruling out a tumbling rifle round) and were definitely larger than .22 caliber. Most gun discharges on occasions like New Year's are handguns, so it could have been any number of handgun rounds in the .35-.40 caliber range. It wasn't large enough to have been 45ACP or .44 Mag. It could have been a .30 caliber rifle round, but given the estimated angle of entry, I would have assumed it would tumble by then. To leave such a clean entry/exit hole would seem unlikely for a tumbling rifle round.

One doesn't refrain from hunting deer with 22LR because the 22LR (at about three times this energy) isn't dangerous, but because it is difficult to get a shot that produces an instant kill with > 90% probability, and one is more likely to chase and possibly lose a deer that is slowly bleeding to death. Hunters that are insensitive to the deer's suffering and willing to give chase will accept a 50% probability and brag about it later.

unless there are buildings, of course.

So no. It can very much still be fatal

For this story you heard... what cartridge was used? What bullet weight? What muzzle velocity? Where was the gun pointed? Do you have any actual information to share?

This doesn't tell us much unfortunately. From Hatcher's testing done in the 1920's he found terminal velocity of a bullet fired straight up which does not destabilize is about 300fps, which with a fairly heavy .30-06 bullet results in about 30ft/lbs of energy. He noted that it was typically accepted that 60ft/lbs of energy was required to produce a disabling wound.

Of course, even at 30ft/lbs is it possible though statistically improbable that a falling bullet could strike in such a way as to cause serious injury or death.

It's very hard to get good numbers on injuries and deaths from this phenomenon, but the evidence points to it being an extraordinarily uncommon occurrence despite the fact that some cities, such as Dallas, report up to 1000 complaints for celebratory gunfire per year.

Note that I am not claiming it is a good idea.

The Chris Dorner fiasco came about because the LAPD fired on civilians they believed to be Dorner, they were incompetent in doing so, but firing on a suspected dangerous gunman is not a bad thing in principle.

But as far as I can tell the cops had no reason to suspect that there was any imminent danger from these flying robots, and were firing in the air just because they didn't like being recorded.

Firing into the air in a civilian area seems to me to be a much bigger issue than them trying to suppress news coverage of protests.

Well, for some very limited values of safely. Actually firing the shotgun would very likely cause return fire directed at the crowd.

So for all the mathematics posts in reply to your post, the real answer is, "Yes, it can happen. Yes, it is a dangerous thing to do"