I dunno anything about guns or bullets, but that gun website says that .223 is "safer for home defense", compared to pistols and shotguns, so says the FBI.
I don't know a huge amount of guns and bullets, but I did grow up around guns in a country where you are legally required to know about gun safety in order to own a gun. The actually important part of that website is when they point out that any gun powerful enough that you might consider using it for "home defence" can easily penetrate a door or interior wall. Guns in general have a lot of penetrating power, that is literally how they work.
So, I actually went and checked the sources on that website. It's interesting. Firstly, I don't get where the FBI comes in. There's a broken link to another gun article published by someone who apparently claimed to be a retired FBI agent, but the article doesn't seem to exist and the name doesn't produce any meaningful search results. Even if it did exist and this person is who they claim to be, that's not the FBI. All the other links are just tests by random gun owners who often openly admit they have little knowledge of ballistics and are generally using a similar setup of spaced drywall sheets to represent the interior of a house. Many of these "tests" are carried out using ammunition specifically created and marketed for its low penetration qualities or high fragmentation, which in many cases don't live up to expectation. In the tests with standard full metal jacket .223 rounds, they typically penetrate all the layers of drywall used in the test
despite tumbling.
So, tumbling happens because when the bullet strikes an object, it forms a bubble of pressure that deforms the front of the bullet and thus creates uneven drag. This makes the bullet tumble. But the bullet doesn't stop tumbling once it exits whatever it was fired at, it continues to tumble through the air, encountering more air resistance and slowing down faster. That is why .223 rounds are relatively good at not penetrating multiple layers of spaced drywall, because between those layers of drywall are large gaps of air. Again though, the sheer amount of energy the bullet starts with is so much greater that it can still penetrate several walls before being stopped completely, but if you live in a big house and you don't want to injure someone three rooms away, .223 may be relatively safer because it will slow down quicker after penetrating the first wall. Again though, the flipside of this is that the trajectory of the bullet becomes erratic once it starts tumbling, and you cannot predict where it will end up.
But this doesn't equate at all to penetration of a human body, because a human body doesn't contain large gaps of air in which the tumbling bullet can slow down. To test how projectiles penetrate a human body, you'd need to do a ballistic gel test. After looking up ballistic gel tests for .223 ammo, I've learned that the amount of penetration actually varies hugely depending on the precise make and composition of the ammunition, with some being similar to pistol ammunition and others being far more penetrating, so in conclusion there is no way to tell from the calibre and model of the firearm how much danger there was of over-penetration.
But, since we're talking about the FBI, let me quote
Handgun Wounding Factors and Effectiveness, an actual FBI report detailing what should be looked for in a handgun cartridge, on the subject of overpenetration of a living body.
An issue that must be addressed is the fear of over penetration widely expressed on the part of law enforcement. The concern that a bullet would pass through the body of a subject and injure an innocent bystander is clearly exaggerated. Any review of law enforcement shootings will reveal that the great majority of shots fired by officers do not hit any subjects at all. It should be obvious that the relatively few shots that do hit a subject are not somehow more dangerous to bystanders than the shots that miss the subject entirely
You cannot safely shoot someone in a crowd without the risk of hitting other people. Even if we put aside the penetration characteristics of the bullet itself, or assume the shooter is deliberately using less penetrating ammunition, it is not possible even for a highly trained person to guarantee all shots will safely hit their target centre mass, let alone some sperm playing dress-up as a cop.