We donāt actually know for certain!
Itās all a gradual series of transitional forms, and where we stick the arbitrary point of ābirds start hereā depends on on what you consider to be their defining traits.
Do we define them by having feathers? If we do that, then fuzzy feather-like structures actually seem to go all the way back to the common ancestor of all dinosaurs and pterosaurs ā which would therefore make every single member of those groups birds too.
If youāre not into the idea of calling things like Pteranodon and Triceratops ābirdsā, then maybe we should go more specific and use the presence of pennaceous feathers instead. The kind with quills and barbs that we normally associate with birdsā wings.
Those are at least found only in theropod dinosaurs, except⦠theyāve been found in ornithomimosaurs, a group well before anything weād usually say was a ābirdā. (And they might go even further back, since there are possible quill knobs known from the carnosaur Concavenator. And then Tyrannosaurus also gets to be a bird!)
Okay, thatās still a bit weird. What about being able to fly?
That at least seems to have happened in Pennaraptora, a group of theropods that includes oviraptorosaurs, dromaeosaurs, Archaeopteryx, and everything else leading to modern birds. So weāre definitely getting much more bird-like in here.
But we donāt know exactly when flight originated. Opinions on whether Archaeopteryx could actually fly go back and forth constantly (the newest research suggests yes) ā but some earlier small dromaeosaurs might also have been capable of powered flight. And thereās even the possibility that bird-like flight independently evolved multiple times in different branches of this particular family tree.
Then thereās always Archaeopteryx as the traditional option for āfirst birdā, but weāve found so many other similar birdy things by now thatās itās not actually unique anymore. Pick out any defining ābirdā feature in Archie and thereās an older dinosaur with the same trait which then also has to be a bird too.
And so we end up with the most conservative option: Neornithes (or Aves), the group that contains all modern birds going back to their last common ancestor. Those are all definitely unquestionably birds, butā¦
ā¦Then things like Confuciusornis and enantiornitheans and Ichthyornis donāt get to be birds. And theyāre very birdy-looking. If you saw one alive youād call it a bird.
But if we keep backtracking through the evolutionary tree along everything that still looks like a bird we just end up right back in the pennaraptoran mess again. So thatās not much help either.
A lot of paleontologists still tend to consider ābirdsā starting at a point just before Archaeopteryx, in a group called Avialae, but thereās also a lot of inconsistency and disagreements. Really you can pick whatever definition you personally like best and just roll with that, because thereās unlikely to be a Definite Official Bird Definition anytime soon.